benchlm September 11, 2026 · llm-stats · json

the llm value index

One number: capability per dollar. 492 models from two public sources. Price is blended $/1M tokens (3 input : 1 output).

value capability coding long context frontier budget best value

pareto frontier other models The 50 most capable models with published pricing. Split lines mark the medians: $2.70 per 1M blended, capability 78.9. Filled points sit on the pareto frontier.

sorted by capability per blended dollar, highest first

blended $ range capability range

	model	capability	blended $/1M	context	value
	Gemma 4 E4B · Google pareto mid open weights	58.6	$0.04	131K	1465
in $ 0.02 · out $ 0.1 · speed - tok/s · sources ls / bl · model id gemma-4-e4b-it · no recent coverage matched
	Llama 3.1 8B Instruct · Meta pareto budget open weights	30.4	$0.025	131K	1216
in $ 0.02 · out $ 0.04 · speed - tok/s · sources ls · model id llama-3.1-8b-instruct · no recent coverage matched
	IBM Granite 4.2 3B · IBM mid open weights	54.8	$0.052	131K	1044
in $ 0.03 · out $ 0.12 · speed - tok/s · sources ls / bl · model id granite-4.2-3b · no recent coverage matched
	Qwen3.7 Flash · Alibaba budget	49.0	$0.055	1M	891
in $ 0.03 · out $ 0.13 · speed - tok/s · sources - / bl · model id qwen3-7-flash · no recent coverage matched
	Mistral Small 3 24B Instruct · Mistral AI budget open weights	45.3	$0.058	33K	788
in $ 0.05 · out $ 0.08 · speed - tok/s · sources ls · model id mistral-small-24b-instruct-2501 · no recent coverage matched
	Qwen3.5-9B · Alibaba Cloud / Qwen Team pareto frontier open weights	81.7	$0.113	262K	726
in $ 0.1 · out $ 0.15 · speed - tok/s · sources ls · model id qwen3.5-9b · no recent coverage matched
	GPT OSS 20B · OpenAI budget open weights	41.2	$0.058	131K	717
in $ 0.03 · out $ 0.14 · speed - tok/s · sources ls / bl · model id gpt-oss-20b · no recent coverage matched
	GPT OSS 120B · OpenAI budget open weights	47.5	$0.07	131K	676
in $ 0.037 · out $ 0.17 · speed 3 tok/s · sources ls / bl · model id gpt-oss-120b · no recent coverage matched
	Phi 4 · Microsoft mid open weights	56.1	$0.087	16K	641
in $ 0.07 · out $ 0.14 · speed - tok/s · sources ls / bl · model id phi-4 · no recent coverage matched
	DeepSeek-V4-Flash-Max · DeepSeek mid open weights	70.7	$0.113	1.0M	628
in $ 0.09 · out $ 0.18 · speed - tok/s · sources ls · model id deepseek-v4-flash-max · no recent coverage matched
	DeepSeek-V4-Flash-0423 · DeepSeek mid open weights	68.8	$0.113	1.0M	612
in $ 0.09 · out $ 0.18 · speed - tok/s · sources ls · model id deepseek-v4-flash-0423 · no recent coverage matched
	Laguna XS 2.1 · Poolside mid open weights	70.9	$0.125	262K	567
in $ 0.1 · out $ 0.2 · speed - tok/s · sources ls / bl · model id laguna-xs-2.1 · no recent coverage matched
	Muse Spark 1.2 · Meta mid	70.5	$0.125	1.0M	564
in $ 0.1 · out $ 0.2 · speed 169 tok/s · sources ls / bl · model id muse-spark-1.2 · no recent coverage matched
	Gemma 3 12B · Google budget open weights	40.9	$0.075	131K	545
in $ 0.05 · out $ 0.15 · speed - tok/s · sources ls · model id gemma-3-12b-it · no recent coverage matched
	Ling 3.0 Flash · InclusionAI budget	47.5	$0.09	131K	528
in $ 0.06 · out $ 0.18 · speed - tok/s · sources ls / bl · model id ling-3.0-flash · no recent coverage matched
	IBM Granite 4.2 8B · IBM mid open weights	55.9	$0.107	131K	520
in $ 0.06 · out $ 0.25 · speed - tok/s · sources ls / bl · model id granite-4.2-8b · no recent coverage matched
	Gemma 3 4B · Google budget open weights	30.8	$0.063	131K	493
in $ 0.05 · out $ 0.1 · speed - tok/s · sources ls · model id gemma-3-4b-it · no recent coverage matched
	Nemotron 3 Nano (30B A3B) · NVIDIA budget open weights	43.1	$0.087	262K	493
in $ 0.05 · out $ 0.2 · speed - tok/s · sources ls / bl · model id nemotron-3-nano-30b-a3b · no recent coverage matched
	Seed 2.0 Mini · ByteDance frontier	79.0	$0.175	256K	451
in $ 0.1 · out $ 0.4 · speed - tok/s · sources ls / bl · model id seed-2.0-mini · no recent coverage matched
	Mistral Small 3.2 24B Instruct · Mistral AI budget open weights	46.1	$0.106	128K	434
in $ 0.075 · out $ 0.2 · speed - tok/s · sources ls · model id mistral-small-3.2-24b-instruct-2506 · no recent coverage matched
	Step-3.5-Flash · StepFun mid open weights	74.4	$0.175	66K	425
in $ 0.1 · out $ 0.4 · speed 243 tok/s · sources ls / bl · model id step-3.5-flash · no recent coverage matched
	Gemma 3 27B · Google budget open weights	42.4	$0.1	131K	424
in $ 0.08 · out $ 0.16 · speed - tok/s · sources ls / bl · model id gemma-3-27b-it · no recent coverage matched
	Nemotron 3.5 Lightning (30B A3B) · NVIDIA budget open weights	46.2	$0.11	262K	420
in $ 0.08 · out $ 0.2 · speed - tok/s · sources ls / bl · model id nemotron-3.5-lightning-30b-a3b · no recent coverage matched
	Llama 4 Scout · Meta mid open weights	57.2	$0.15	328K	381
in $ 0.1 · out $ 0.3 · speed - tok/s · sources ls / bl · model id llama-4-scout · no recent coverage matched
	Qwen3-235B-A22B-Instruct-2507 · Alibaba Cloud / Qwen Team frontier open weights	77.5	$0.205	262K	378
in $ 0.09 · out $ 0.55 · speed - tok/s · sources ls · model id qwen3-235b-a22b-instruct-2507 · no recent coverage matched
	Ministral 3 3B (Reasoning) · Mistral budget open weights	37.6	$0.1	256K	376
in $ 0.1 · out $ 0.1 · speed - tok/s · sources - / bl · model id ministral-3-3b-reasoning · no recent coverage matched
	Gemma 4 31B · Google mid open weights	55.4	$0.152	262K	363
in $ 0.09 · out $ 0.34 · speed 50 tok/s · sources ls / bl · model id gemma-4-31b-it · no recent coverage matched
	Gemma 4 26B-A4B · Google budget open weights	49.8	$0.138	262K	362
in $ 0.07 · out $ 0.34 · speed 103 tok/s · sources ls / bl · model id gemma-4-26b-a4b-it · no recent coverage matched
	Qwen3 30B A3B · Alibaba Cloud / Qwen Team mid open weights	65.8	$0.185	128K	356
in $ 0.1 · out $ 0.44 · speed - tok/s · sources ls · model id qwen3-30b-a3b · no recent coverage matched
	LongCat-Flash-Lite · Meituan mid open weights	60.6	$0.175	256K	346
in $ 0.1 · out $ 0.4 · speed - tok/s · sources ls / bl · model id longcat-flash-lite · no recent coverage matched
	Hy3 · Tencent pareto frontier open weights	84.2	$0.25	262K	337
in $ 0.14 · out $ 0.58 · speed - tok/s · sources ls / bl · model id hy3 · no recent coverage matched
	Llama 3.3 70B Instruct · Meta mid open weights	50.5	$0.155	131K	326
in $ 0.1 · out $ 0.32 · speed - tok/s · sources ls · model id llama-3.3-70b-instruct · no recent coverage matched
	Nemotron 3 Super (120B A12B) · NVIDIA mid open weights	53.1	$0.164	262K	324
in $ 0.085 · out $ 0.4 · speed - tok/s · sources ls / bl · model id nemotron-3-super-120b-a12b · no recent coverage matched
	Qwen3.5 Flash · Alibaba mid	56.1	$0.175	1M	321
in $ 0.1 · out $ 0.4 · speed - tok/s · sources - / bl · model id qwen3-5-flash · no recent coverage matched
	GPT-5 nano · OpenAI budget	40.0	$0.138	400K	291
in $ 0.05 · out $ 0.4 · speed - tok/s · sources ls / bl · model id gpt-5-nano-2025-08-07 · no recent coverage matched
	GPT-4.1 nano · OpenAI mid	50.3	$0.175	1.0M	287
in $ 0.1 · out $ 0.4 · speed 236 tok/s · sources ls / bl · model id gpt-4.1-nano-2025-04-14 · no recent coverage matched
	Qwen3.8 Flash · Alibaba Cloud / Qwen Team mid	63.8	$0.23	1M	277
in $ 0.15 · out $ 0.47 · speed 74 tok/s · sources ls · model id qwen3.8-flash · no recent coverage matched
	MiMo-V2.5 · Xiaomi mid open weights	57.9	$0.21	1.0M	276
in $ 0.168 · out $ 0.336 · speed - tok/s · sources ls / bl · model id mimo-v2.5 · no recent coverage matched
	Mistral Small 4 · Mistral AI mid open weights	71.2	$0.263	256K	271
in $ 0.15 · out $ 0.6 · speed 177 tok/s · sources ls / bl · model id mistral-small-latest · no recent coverage matched
	Qwen3 VL 30B A3B Instruct · Alibaba Cloud / Qwen Team mid open weights	70.4	$0.263	262K	268
in $ 0.15 · out $ 0.6 · speed - tok/s · sources ls · model id qwen3-vl-30b-a3b-instruct · no recent coverage matched
	Ministral 3 8B (Reasoning) · Mistral budget open weights	38.4	$0.15	256K	256
in $ 0.15 · out $ 0.15 · speed - tok/s · sources - / bl · model id ministral-3-8b-reasoning · no recent coverage matched
	Grok-3 Mini · xAI frontier	84.0	$0.35	128K	240
in $ 0.3 · out $ 0.5 · speed - tok/s · sources ls / bl · model id grok-3-mini · no recent coverage matched
	Ministral 3 14B (Reasoning) · Mistral budget open weights	47.2	$0.2	256K	236
in $ 0.2 · out $ 0.2 · speed - tok/s · sources - / bl · model id ministral-3-14b-reasoning · no recent coverage matched
	GLM-5.3-Flash · Zhipu AI mid open weights	55.3	$0.237	1.0M	233
in $ 0.15 · out $ 0.5 · speed 81 tok/s · sources ls / bl · model id glm-5.3-flash · news ↗
	DeepSeek-V3.2 · DeepSeek mid open weights	65.4	$0.29	164K	226
in $ 0.26 · out $ 0.38 · speed - tok/s · sources ls · model id deepseek-v3.2 · news ↗
	IBM Granite 4.2 30B · IBM mid open weights	61.7	$0.282	131K	218
in $ 0.16 · out $ 0.65 · speed - tok/s · sources ls / bl · model id granite-4.2-30b · no recent coverage matched
	Qwen3-Next-80B-A3B-Instruct · Alibaba Cloud / Qwen Team mid open weights	72.9	$0.343	262K	213
in $ 0.09 · out $ 1.10 · speed - tok/s · sources ls · model id qwen3-next-80b-a3b-instruct · no recent coverage matched
	GPT-5.6 Luna · OpenAI pareto frontier	92.3	$0.45	1.1M	205
in $ 0.2 · out $ 1.20 · speed 141 tok/s · sources ls / bl · model id gpt-5.6-luna · news ↗
	Llama 4 Maverick · Meta mid open weights	69.8	$0.35	1.0M	199
in $ 0.2 · out $ 0.8 · speed - tok/s · sources ls / bl · model id llama-4-maverick · no recent coverage matched
	Mercury 2 · Inception mid	74.0	$0.375	128K	197
in $ 0.25 · out $ 0.75 · speed - tok/s · sources ls / bl · model id mercury-2 · no recent coverage matched
	DeepSeek-V4.1-Flash · DeepSeek mid open weights	63.9	$0.33	1.0M	194
in $ 0.22 · out $ 0.66 · speed - tok/s · sources ls / bl · model id deepseek-v4.1-flash · no recent coverage matched
	Gemini 2.5 Flash-Lite · Google budget open weights	33.8	$0.175	1M	193
in $ 0.1 · out $ 0.4 · speed - tok/s · sources ls / bl · model id gemini-2.5-flash-lite · no recent coverage matched
	Qwen3.6-35B-A3B · Alibaba Cloud / Qwen Team mid open weights	60.3	$0.313	262K	193
in $ 0.1 · out $ 0.95 · speed - tok/s · sources ls · model id qwen3.6-35b-a3b · no recent coverage matched
	Qwen3.5-35B-A3B · Alibaba Cloud / Qwen Team mid open weights	66.9	$0.355	262K	189
in $ 0.14 · out $ 1.00 · speed - tok/s · sources ls / bl · model id qwen3.5-35b-a3b · no recent coverage matched
	DeepSeek V3.2 · DeepSeek mid open weights	56.9	$0.315	128K	181
in $ 0.28 · out $ 0.42 · speed - tok/s · sources - / bl · model id deepseek-v3-2 · no recent coverage matched
	Grok 4.1 Fast · xAI budget	47.3	$0.275	2M	172
in $ 0.2 · out $ 0.5 · speed - tok/s · sources - / bl · model id grok-4-1-fast · no recent coverage matched
	DeepSeek-V3 0324 · DeepSeek mid open weights	68.4	$0.405	164K	169
in $ 0.24 · out $ 0.9 · speed - tok/s · sources ls · model id deepseek-v3-0324 · no recent coverage matched
	MiniMax M3 · MiniMax frontier open weights	80.5	$0.485	1M	166
in $ 0.28 · out $ 1.10 · speed 109 tok/s · sources ls / bl · model id minimax-m3 · no recent coverage matched
	MAI-Code-1.1-Flash · Microsoft mid	72.6	$0.45	256K	161
in $ 0.2 · out $ 1.20 · speed - tok/s · sources ls / bl · model id mai-code-1.1-flash · no recent coverage matched
	Ministral 3 3B · Mistral budget open weights	16.0	$0.1	256K	160
in $ 0.1 · out $ 0.1 · speed - tok/s · sources - / bl · model id ministral-3-3b · no recent coverage matched
	MiniMax M2.5 · MiniMax frontier open weights	80.2	$0.525	1M	153
in $ 0.3 · out $ 1.20 · speed 147 tok/s · sources ls / bl · model id minimax-m2.5 · no recent coverage matched
	Solar Pro 4 · Upstage frontier	79.8	$0.525	524K	152
in $ 0.3 · out $ 1.20 · speed - tok/s · sources ls / bl · model id solar-pro4 · no recent coverage matched
	Ministral 3 14B · Mistral budget open weights	30.3	$0.2	256K	152
in $ 0.2 · out $ 0.2 · speed - tok/s · sources - / bl · model id ministral-3-14b · no recent coverage matched
	Qwen3-Coder 480B A35B Instruct · Alibaba Cloud / Qwen Team mid open weights	69.6	$0.475	262K	147
in $ 0.3 · out $ 1.00 · speed - tok/s · sources ls · model id qwen3-coder-480b-a35b-instruct · no recent coverage matched
	Grok Code Fast 1 · xAI mid	70.8	$0.525	256K	135
in $ 0.2 · out $ 1.50 · speed - tok/s · sources ls / bl · model id grok-code-fast-1 · no recent coverage matched
	Qwen2.5 72B Instruct · Alibaba Cloud / Qwen Team budget open weights	49.0	$0.37	33K	132
in $ 0.36 · out $ 0.4 · speed - tok/s · sources ls / bl · model id qwen-2.5-72b-instruct · no recent coverage matched
	Inkling-Small · Thinking Machines Lab mid open weights	67.1	$0.525	524K	128
in $ 0.3 · out $ 1.20 · speed - tok/s · sources ls / bl · model id inkling-small · news ↗
	DeepSeek-V3.1 · DeepSeek mid open weights	52.3	$0.425	164K	123
in $ 0.25 · out $ 0.95 · speed - tok/s · sources ls / bl · model id deepseek-v3.1 · no recent coverage matched
	Trinity-Large-Preview · Arcee AI mid open weights	53.7	$0.438	512K	123
in $ 0.25 · out $ 1.00 · speed - tok/s · sources - / bl · model id trinity-large-preview · no recent coverage matched
	Ministral 3 8B · Mistral budget open weights	17.9	$0.15	256K	119
in $ 0.15 · out $ 0.15 · speed - tok/s · sources - / bl · model id ministral-3-8b · no recent coverage matched
	GPT-5.4 nano · OpenAI mid	53.6	$0.463	400K	116
in $ 0.2 · out $ 1.25 · speed 162 tok/s · sources ls / bl · model id gpt-5.4-nano · no recent coverage matched
	Muse Glimmer-30B · Meta mid open weights	60.5	$0.525	131K	115
in $ 0.3 · out $ 1.20 · speed - tok/s · sources ls / bl · model id muse-glimmer-30b · no recent coverage matched
	Step 3.7 Flash · StepFun mid open weights	50.0	$0.438	262K	114
in $ 0.2 · out $ 1.15 · speed - tok/s · sources ls / bl · model id step-3.7-flash · no recent coverage matched
	GLM-4.5-Air · Zhipu AI budget open weights	47.7	$0.425	128K	112
in $ 0.2 · out $ 1.10 · speed - tok/s · sources ls / bl · model id glm-4.5-air · no recent coverage matched
	DeepSeek-V4-Pro-0813 · DeepSeek mid open weights	60.0	$0.544	1.0M	110
in $ 0.435 · out $ 0.87 · speed 65 tok/s · sources ls / bl · model id deepseek-v4-pro-0813 · no recent coverage matched
	MiMo-V2.5-Pro · Xiaomi mid open weights	59.9	$0.544	1.0M	110
in $ 0.435 · out $ 0.87 · speed - tok/s · sources ls / bl · model id mimo-v2.5-pro · no recent coverage matched
	DeepSeek-V3 · DeepSeek mid open weights	50.6	$0.463	164K	109
in $ 0.32 · out $ 0.89 · speed - tok/s · sources ls / bl · model id deepseek-v3 · no recent coverage matched
	MiniMax M2.1 · MiniMax mid open weights	56.7	$0.525	1M	108
in $ 0.3 · out $ 1.20 · speed - tok/s · sources ls · model id minimax-m2.1 · no recent coverage matched
	Seed 1.8 · ByteDance mid	72.9	$0.688	256K	106
in $ 0.25 · out $ 2.00 · speed - tok/s · sources ls · model id seed-1.8 · no recent coverage matched
	Trinity-Large-Thinking · Arcee AI budget open weights	43.4	$0.412	512K	105
in $ 0.25 · out $ 0.9 · speed - tok/s · sources - / bl · model id trinity-large-thinking · no recent coverage matched
	MiniMax M2.7 · MiniMax mid open weights	55.2	$0.525	205K	105
in $ 0.3 · out $ 1.20 · speed - tok/s · sources ls / bl · model id minimax-m2.7 · no recent coverage matched
	Llama 3.1 70B Instruct · Meta budget open weights	41.7	$0.4	131K	104
in $ 0.4 · out $ 0.4 · speed - tok/s · sources ls · model id llama-3.1-70b-instruct · no recent coverage matched
	MiniMax M2 · MiniMax mid open weights	53.3	$0.525	1M	102
in $ 0.3 · out $ 1.20 · speed - tok/s · sources ls · model id minimax-m2 · no recent coverage matched
	Hermes 3 70B · Nous Research mid open weights	66.1	$0.7	131K	94.4
in $ 0.7 · out $ 0.7 · speed - tok/s · sources ls · model id hermes-3-70b · no recent coverage matched
	GPT-4o mini · OpenAI budget	24.5	$0.263	128K	93.3
in $ 0.15 · out $ 0.6 · speed - tok/s · sources ls / bl · model id gpt-4o-mini-2024-07-18 · no recent coverage matched
	Gemini 3.1 Flash-Lite · Google mid	51.5	$0.563	1M	91.6
in $ 0.25 · out $ 1.50 · speed - tok/s · sources ls / bl · model id gemini-3.1-flash-lite-preview · no recent coverage matched
	GLM-4.7 · Zhipu AI mid open weights	67.4	$0.738	203K	91.4
in $ 0.4 · out $ 1.75 · speed - tok/s · sources ls / bl · model id glm-4.7 · no recent coverage matched
	Nova 2 Lite · Amazon mid	72.1	$0.85	1M	84.8
in $ 0.3 · out $ 2.50 · speed - tok/s · sources ls · model id nova-2-lite · no recent coverage matched
	Qwen3.5-122B-A10B · Alibaba Cloud / Qwen Team mid open weights	68.7	$0.818	262K	84.0
in $ 0.29 · out $ 2.40 · speed - tok/s · sources ls / bl · model id qwen3.5-122b-a10b · no recent coverage matched
	Qwen3.5-27B · Alibaba Cloud / Qwen Team mid open weights	68.8	$0.845	262K	81.4
in $ 0.26 · out $ 2.60 · speed 24 tok/s · sources ls / bl · model id qwen3.5-27b · no recent coverage matched
	Seed 2.0 Pro · ByteDance frontier	82.7	$1.13	256K	73.5
in $ 0.5 · out $ 3.00 · speed 30 tok/s · sources ls · model id seed-2.0-pro · no recent coverage matched
	GPT-5 mini · OpenAI budget	49.5	$0.688	400K	72.0
in $ 0.25 · out $ 2.00 · speed 352 tok/s · sources ls / bl · model id gpt-5-mini-2025-08-07 · no recent coverage matched
	Nemotron 3 Ultra (550B A55B) · NVIDIA mid open weights	65.0	$0.925	262K	70.3
in $ 0.5 · out $ 2.20 · speed - tok/s · sources ls / bl · model id nemotron-3-ultra-550b-a55b · no recent coverage matched
	Gemini 3.5 Flash-Lite · Google mid	59.0	$0.85	1.0M	69.4
in $ 0.3 · out $ 2.50 · speed 223 tok/s · sources ls / bl · model id gemini-3.5-flash-lite · no recent coverage matched
	Claude 3 Haiku · Anthropic budget	33.3	$0.5	200K	66.6
in $ 0.25 · out $ 1.25 · speed - tok/s · sources ls / bl · model id claude-3-haiku-20240307 · no recent coverage matched
	GLM-4.6 · Zhipu AI mid open weights	55.4	$0.875	203K	63.3
in $ 0.5 · out $ 2.00 · speed - tok/s · sources ls · model id glm-4.6 · no recent coverage matched
	GLM-5.2 · Zhipu AI mid open weights	73.0	$1.16	1.0M	62.8
in $ 0.75 · out $ 2.40 · speed - tok/s · sources ls / bl · model id glm-5.2 · news ↗
	Gemini 3 Flash · Google mid	70.6	$1.13	1M	62.8
in $ 0.5 · out $ 3.00 · speed 320 tok/s · sources ls / bl · model id gemini-3-flash-preview · no recent coverage matched
	Mistral Large 3 · Mistral budget	45.6	$0.75	256K	60.8
in $ 0.5 · out $ 1.50 · speed - tok/s · sources - / bl · model id mistral-large-3 · no recent coverage matched
	Qwen3.6-27B · Alibaba Cloud / Qwen Team mid open weights	63.0	$1.04	262K	60.6
in $ 0.32 · out $ 3.20 · speed 1 tok/s · sources ls / bl · model id qwen3.6-27b · no recent coverage matched
	Gemini 2.5 Flash · Google mid	51.4	$0.85	1M	60.5
in $ 0.3 · out $ 2.50 · speed - tok/s · sources ls / bl · model id gemini-2.5-flash · no recent coverage matched
	Kimi K2.5 · Moonshot AI mid open weights	71.5	$1.20	256K	59.6
in $ 0.6 · out $ 3.00 · speed - tok/s · sources ls / bl · model id kimi-k2.5 · no recent coverage matched
	Qwen3.5-397B-A17B · Alibaba Cloud / Qwen Team mid open weights	64.5	$1.09	262K	59.3
in $ 0.45 · out $ 3.00 · speed - tok/s · sources ls / bl · model id qwen3.5-397b-a17b · no recent coverage matched
	Qwen3.6 Plus · Alibaba Cloud / Qwen Team mid	66.0	$1.13	1M	58.7
in $ 0.5 · out $ 3.00 · speed 111 tok/s · sources ls / bl · model id qwen3.6-plus · no recent coverage matched
	Mistral Large 3 (675B Instruct 2512) · Mistral AI budget open weights	43.9	$0.75	262K	58.5
in $ 0.5 · out $ 1.50 · speed - tok/s · sources ls · model id mistral-large-latest · no recent coverage matched
	DeepSeek V3.2 (Thinking) · DeepSeek mid open weights	55.8	$0.96	128K	58.1
in $ 0.55 · out $ 2.19 · speed - tok/s · sources - / bl · model id deepseek-v3-2-thinking · no recent coverage matched
	Qwen3.8-27B · Alibaba Cloud / Qwen Team mid open weights	60.0	$1.05	262K	57.1
in $ 0.4 · out $ 3.00 · speed - tok/s · sources ls / bl · model id qwen3.8-27b · no recent coverage matched
	Qwen3.5 Plus · Alibaba budget	49.9	$0.9	1M	55.4
in $ 0.4 · out $ 2.40 · speed - tok/s · sources - / bl · model id qwen3-5-plus · no recent coverage matched
	GLM-4.5 · Zhipu AI mid open weights	52.6	$1.00	128K	52.6
in $ 0.6 · out $ 2.20 · speed - tok/s · sources ls / bl · model id glm-4.5 · no recent coverage matched
	DeepSeek-R1-0528 · DeepSeek budget open weights	47.8	$0.912	164K	52.4
in $ 0.5 · out $ 2.15 · speed - tok/s · sources ls · model id deepseek-r1-0528 · no recent coverage matched
	DeepSeek-R1 · DeepSeek mid open weights	50.2	$0.96	128K	52.3
in $ 0.55 · out $ 2.19 · speed - tok/s · sources ls / bl · model id deepseek-r1 · no recent coverage matched
	Gemini 3.8 Flash · Google frontier	75.6	$1.50	1.0M	50.4
in $ 0.75 · out $ 3.75 · speed 224 tok/s · sources ls / bl · model id gemini-3.8-flash · news ↗
	GLM-5 · Zhipu AI frontier open weights	77.8	$1.55	200K	50.2
in $ 1.00 · out $ 3.20 · speed - tok/s · sources ls / bl · model id glm-5 · no recent coverage matched
	Kimi K2.7 Code · Moonshot AI mid open weights	65.5	$1.36	262K	48.2
in $ 0.68 · out $ 3.40 · speed - tok/s · sources ls / bl · model id kimi-k2.7-code · no recent coverage matched
	Kimi K2.6 · Moonshot AI mid open weights	69.0	$1.44	262K	48.0
in $ 0.75 · out $ 3.50 · speed 87 tok/s · sources ls · model id kimi-k2.6 · no recent coverage matched
	Kimi K2.5 (Reasoning) · Moonshot AI mid	57.0	$1.20	256K	47.5
in $ 0.6 · out $ 3.00 · speed - tok/s · sources - / bl · model id kimi-k2-5-reasoning · no recent coverage matched
	Aion-2.0 · Aion Labs budget	46.5	$1.00	128K	46.5
in $ 0.8 · out $ 1.60 · speed - tok/s · sources - / bl · model id aion-2-0 · no recent coverage matched
	Gemini 3.7 Flash · Google mid	68.5	$1.50	1.0M	45.7
in $ 0.75 · out $ 3.75 · speed 235 tok/s · sources ls / bl · model id gemini-3.7-flash · news ↗
	Inkling · Thinking Machines Lab frontier open weights	77.6	$1.73	524K	45.0
in $ 0.95 · out $ 4.05 · speed - tok/s · sources ls / bl · model id inkling · no recent coverage matched
	DeepSeek-V4-Pro-Max · DeepSeek mid open weights	73.0	$1.63	1.0M	44.9
in $ 1.30 · out $ 2.60 · speed - tok/s · sources ls · model id deepseek-v4-pro-max · no recent coverage matched
	GPT-4.1 mini · OpenAI budget	30.8	$0.7	1.0M	44.0
in $ 0.4 · out $ 1.60 · speed 333 tok/s · sources ls / bl · model id gpt-4.1-mini-2025-04-14 · no recent coverage matched
	Qwen3.5 397B (Reasoning) · Alibaba mid open weights	57.2	$1.35	128K	42.4
in $ 0.6 · out $ 3.60 · speed - tok/s · sources - / bl · model id qwen3-5-397b-reasoning · no recent coverage matched
	GLM-5.1 · Zhipu AI mid open weights	69.3	$1.66	203K	41.7
in $ 1.05 · out $ 3.50 · speed - tok/s · sources ls / bl · model id glm-5.1 · news ↗
	GPT-3.5 Turbo · OpenAI budget	30.8	$0.75	16K	41.1
in $ 0.5 · out $ 1.50 · speed - tok/s · sources ls · model id gpt-3.5-turbo-0125 · no recent coverage matched
	Kimi 2.6 · Moonshot AI mid	65.4	$1.71	256K	38.2
in $ 0.95 · out $ 4.00 · speed - tok/s · sources - / bl · model id kimi-2-6 · no recent coverage matched
	Grok 4.3 · xAI mid	59.7	$1.56	1M	38.2
in $ 1.25 · out $ 2.50 · speed - tok/s · sources ls / bl · model id grok-4.3 · no recent coverage matched
	Qwen3.7 Max · Alibaba Cloud / Qwen Team mid	71.4	$1.88	1M	38.1
in $ 1.25 · out $ 3.75 · speed - tok/s · sources ls / bl · model id qwen3.7-max · no recent coverage matched
	GLM-5 (Reasoning) · Z.AI mid open weights	57.4	$1.55	200K	37.0
in $ 1.00 · out $ 3.20 · speed - tok/s · sources - / bl · model id glm-5-reasoning · no recent coverage matched
	Claude Haiku 4.5 · Anthropic mid	73.2	$2.00	200K	36.6
in $ 1.00 · out $ 5.00 · speed 57 tok/s · sources ls / bl · model id claude-haiku-4-5-20251001 · no recent coverage matched
	GPT-5.4 mini · OpenAI mid	58.1	$1.69	400K	34.4
in $ 0.75 · out $ 4.50 · speed 165 tok/s · sources ls / bl · model id gpt-5.4-mini · no recent coverage matched
	Qwen3 Max Thinking · Alibaba Cloud / Qwen Team frontier	81.4	$2.40	256K	33.9
in $ 1.20 · out $ 6.00 · speed - tok/s · sources ls · model id qwen3-max-thinking · no recent coverage matched
	o3-mini · OpenAI mid	63.3	$1.93	200K	32.9
in $ 1.10 · out $ 4.40 · speed - tok/s · sources ls / bl · model id o3-mini · no recent coverage matched
	GLM-5.3 · Zhipu AI mid open weights	62.5	$1.90	1.0M	32.9
in $ 1.20 · out $ 4.00 · speed 147 tok/s · sources ls / bl · model id glm-5.3 · news ↗
	GLM-5-Turbo · Z.AI mid	61.6	$1.90	200K	32.4
in $ 1.20 · out $ 4.00 · speed - tok/s · sources - / bl · model id glm-5-turbo · no recent coverage matched
	Muse Spark 1.1 · Meta mid	62.1	$2.00	1.0M	31.1
in $ 1.25 · out $ 4.25 · speed 184 tok/s · sources ls · model id muse-spark-1.1 · news ↗
	Grok 4.5 · xAI pareto frontier	93.0	$3.00	500K	31.0
in $ 2.00 · out $ 6.00 · speed 120 tok/s · sources ls / bl · model id grok-4.5 · news ↗
	GLM-5V-Turbo · Zhipu AI mid	57.6	$1.90	200K	30.3
in $ 1.20 · out $ 4.00 · speed - tok/s · sources ls / bl · model id glm-5v-turbo · no recent coverage matched
	o4-mini · OpenAI mid	54.7	$1.93	200K	28.4
in $ 1.10 · out $ 4.40 · speed - tok/s · sources ls / bl · model id o4-mini · no recent coverage matched
	Qwen3.8 Max · Alibaba Cloud / Qwen Team mid open weights	68.1	$2.48	1.0M	27.5
in $ 1.65 · out $ 4.95 · speed 114 tok/s · sources ls / bl · model id qwen3.8-max · news ↗
	Qwen3 Max · Alibaba Cloud / Qwen Team mid	65.8	$2.40	256K	27.4
in $ 1.20 · out $ 6.00 · speed - tok/s · sources ls · model id qwen3-max · no recent coverage matched
	Gemini 1.5 Pro · Google mid	59.1	$2.19	1M	27.0
in $ 1.25 · out $ 5.00 · speed - tok/s · sources ls / bl · model id gemini-1.5-pro · no recent coverage matched
	Mistral Medium 3.5 · Mistral AI frontier open weights	77.6	$3.00	256K	25.9
in $ 1.50 · out $ 7.50 · speed 12 tok/s · sources ls / bl · model id mistral-medium-3-5 · no recent coverage matched
	GPT-5 High · OpenAI frontier	87.3	$3.44	400K	25.4
in $ 1.25 · out $ 10.00 · speed - tok/s · sources ls / bl · model id gpt-5-high-2025-08-07 · no recent coverage matched
	GPT-5.1 · OpenAI frontier	82.2	$3.44	400K	23.9
in $ 1.25 · out $ 10.00 · speed - tok/s · sources ls / bl · model id gpt-5.1-2025-11-13 · no recent coverage matched
	GPT-5.1 Instant · OpenAI frontier	82.2	$3.44	400K	23.9
in $ 1.25 · out $ 10.00 · speed 139 tok/s · sources ls · model id gpt-5.1-instant-2025-11-12 · no recent coverage matched
	Grok 4.6 · xAI mid	70.1	$3.00	500K	23.4
in $ 2.00 · out $ 6.00 · speed 61 tok/s · sources ls / bl · model id grok-4.6 · no recent coverage matched
	Gemini 3.6 Flash · Google mid	68.4	$3.00	1.0M	22.8
in $ 1.50 · out $ 7.50 · speed 360 tok/s · sources ls / bl · model id gemini-3.6-flash · news ↗
	Grok 4.20 · xAI mid	67.1	$3.00	2M	22.4
in $ 2.00 · out $ 6.00 · speed - tok/s · sources - / bl · model id grok-4-20-beta · no recent coverage matched
	Kimi K2 · Moonshot AI budget	23.3	$1.07	128K	21.7
in $ 0.6 · out $ 2.50 · speed - tok/s · sources - / bl · model id kimi-k2 · no recent coverage matched
	GPT-5.2 Instant · OpenAI mid	56.7	$2.63	128K	21.6
in $ 1.50 · out $ 6.00 · speed - tok/s · sources - / bl · model id gpt-5-2-instant · no recent coverage matched
	GPT-5.1 Codex · OpenAI mid	73.7	$3.44	400K	21.4
in $ 1.25 · out $ 10.00 · speed - tok/s · sources ls / bl · model id gpt-5.1-codex · no recent coverage matched
	GPT-5.6 Terra · OpenAI frontier	92.9	$4.50	1.1M	20.6
in $ 2.00 · out $ 12.00 · speed 135 tok/s · sources ls / bl · model id gpt-5.6-terra · no recent coverage matched
	Claude Sonnet 5 · Anthropic mid	71.3	$4.00	1M	17.8
in $ 2.00 · out $ 10.00 · speed - tok/s · sources ls / bl · model id claude-sonnet-5 · news ↗
	Gemini 3.1 Pro · Google frontier	75.4	$4.50	1.0M	16.8
in $ 2.00 · out $ 12.00 · speed - tok/s · sources ls / bl · model id gemini-3.1-pro-preview · no recent coverage matched
	o3 · OpenAI mid	55.7	$3.50	200K	15.9
in $ 2.00 · out $ 8.00 · speed - tok/s · sources ls / bl · model id o3-2025-04-16 · no recent coverage matched
	Gemini 2.5 Pro · Google mid	54.7	$3.44	1.0M	15.9
in $ 1.25 · out $ 10.00 · speed - tok/s · sources ls / bl · model id gemini-2.5-pro · no recent coverage matched
	Gemini 3 Pro · Google mid	71.3	$4.50	2M	15.8
in $ 2.00 · out $ 12.00 · speed - tok/s · sources ls / bl · model id gemini-3-pro-preview · no recent coverage matched
	GPT-5.1-Codex-Max · OpenAI mid	52.2	$3.44	400K	15.2
in $ 1.25 · out $ 10.00 · speed - tok/s · sources - / bl · model id gpt-5-1-codex-max · no recent coverage matched
	GPT-5.2 · OpenAI mid	69.0	$4.81	400K	14.3
in $ 1.75 · out $ 14.00 · speed - tok/s · sources ls / bl · model id gpt-5.2-2025-12-11 · no recent coverage matched
	Grok-3 · xAI frontier	84.6	$6.00	128K	14.1
in $ 3.00 · out $ 15.00 · speed - tok/s · sources ls / bl · model id grok-3 · no recent coverage matched
	Claude Sonnet 4.5 · Anthropic frontier	83.4	$6.00	200K	13.9
in $ 3.00 · out $ 15.00 · speed - tok/s · sources ls / bl · model id claude-sonnet-4-5-20250929 · no recent coverage matched
	GPT-5.3 Codex · OpenAI mid	64.3	$4.81	400K	13.4
in $ 1.75 · out $ 14.00 · speed - tok/s · sources ls / bl · model id gpt-5.3-codex · no recent coverage matched
	Kimi K3 · Moonshot AI mid open weights	74.8	$5.70	1.0M	13.1
in $ 2.85 · out $ 14.25 · speed 3 tok/s · sources ls / bl · model id kimi-k3 · news ↗
	GPT-4o · OpenAI mid	53.6	$4.38	128K	12.3
in $ 2.50 · out $ 10.00 · speed - tok/s · sources ls · model id gpt-4o-2024-05-13 · news ↗
	Claude Sonnet 4.6 · Anthropic mid	72.8	$6.00	1M	12.1
in $ 3.00 · out $ 15.00 · speed - tok/s · sources ls / bl · model id claude-sonnet-4-6 · news ↗
	GPT-4.1 · OpenAI budget	42.1	$3.50	1.0M	12.0
in $ 2.00 · out $ 8.00 · speed - tok/s · sources ls / bl · model id gpt-4.1-2025-04-14 · no recent coverage matched
	Gemini 3.5 Flash · Google budget	40.2	$3.38	1.0M	11.9
in $ 1.50 · out $ 9.00 · speed 497 tok/s · sources ls / bl · model id gemini-3.5-flash · news ↗
	GPT-5.4 · OpenAI mid	66.3	$5.63	1M	11.8
in $ 2.50 · out $ 15.00 · speed - tok/s · sources ls / bl · model id gpt-5.4 · news ↗
	GPT-5.3 Instant · OpenAI mid	56.6	$4.81	128K	11.8
in $ 1.75 · out $ 14.00 · speed - tok/s · sources - / bl · model id gpt-5-3-instant · no recent coverage matched
	GPT-5.2 Codex · OpenAI mid	54.2	$4.81	400K	11.3
in $ 1.75 · out $ 14.00 · speed - tok/s · sources ls / bl · model id gpt-5.2-codex · no recent coverage matched
	Mistral Medium 3.5 128B · Mistral budget open weights	30.1	$3.00	256K	10.0
in $ 1.50 · out $ 7.50 · speed - tok/s · sources - / bl · model id mistral-medium-3-5-128b · no recent coverage matched
	Command A+ · Cohere budget open weights	42.6	$4.38	128K	9.7
in $ 2.50 · out $ 10.00 · speed - tok/s · sources ls / bl · model id command-a-plus-05-2026 · no recent coverage matched
	GPT-4o · OpenAI budget	38.9	$4.38	128K	8.9
in $ 2.50 · out $ 10.00 · speed - tok/s · sources - / bl · model id gpt-4o · news ↗
	GPT-5.6 Sol · OpenAI pareto frontier	94.6	$11.25	1.1M	8.4
in $ 5.00 · out $ 30.00 · speed 85 tok/s · sources ls / bl · model id gpt-5.6-sol · news ↗
	Claude Opus 4.5 · Anthropic frontier	84.0	$10.00	200K	8.4
in $ 5.00 · out $ 25.00 · speed - tok/s · sources ls / bl · model id claude-opus-4-5-20251101 · no recent coverage matched
	GPT-4o · OpenAI budget	36.2	$4.38	128K	8.3
in $ 2.50 · out $ 10.00 · speed - tok/s · sources ls · model id gpt-4o-2024-08-06 · news ↗
	Claude Opus 4.8 · Anthropic frontier	80.0	$10.00	1M	8.0
in $ 5.00 · out $ 25.00 · speed 59 tok/s · sources ls / bl · model id claude-opus-4-8 · no recent coverage matched
	Claude Opus 4.7 · Anthropic frontier	78.8	$10.00	1M	7.9
in $ 5.00 · out $ 25.00 · speed - tok/s · sources ls / bl · model id claude-opus-4-7 · no recent coverage matched
	GPT-5.5 Instant · OpenAI frontier	85.6	$11.25	400K	7.6
in $ 5.00 · out $ 30.00 · speed - tok/s · sources ls · model id gpt-5.5-instant · news ↗
	Claude Opus 4.6 · Anthropic frontier	75.1	$10.00	1M	7.5
in $ 5.00 · out $ 25.00 · speed 34 tok/s · sources ls / bl · model id claude-opus-4-6 · news ↗
	Claude Opus 4.7 (Adaptive) · Anthropic mid	69.1	$10.00	1M	6.9
in $ 5.00 · out $ 25.00 · speed - tok/s · sources - / bl · model id claude-opus-4-7-adaptive · no recent coverage matched
	Claude 4 Sonnet · Anthropic budget	39.4	$6.00	200K	6.6
in $ 3.00 · out $ 15.00 · speed - tok/s · sources - / bl · model id claude-4-sonnet · no recent coverage matched
	GPT-5.5 · OpenAI mid	72.9	$11.25	1.1M	6.5
in $ 5.00 · out $ 30.00 · speed - tok/s · sources ls / bl · model id gpt-5.5 · news ↗
	Claude Opus 5 · Anthropic mid	64.7	$10.00	1M	6.5
in $ 5.00 · out $ 25.00 · speed 55 tok/s · sources ls / bl · model id claude-opus-5 · news ↗
	Claude 3.5 Sonnet · Anthropic budget	32.0	$6.00	200K	5.3
in $ 3.00 · out $ 15.00 · speed - tok/s · sources - / bl · model id claude-3-5-sonnet · no recent coverage matched
	GPT-6 Astra · OpenAI pareto frontier	96.0	$20.00	1.1M	4.8
in $ 10.00 · out $ 50.00 · speed 93 tok/s · sources ls / bl · model id gpt-6-astra · news ↗
	Claude Fable 5 · Anthropic frontier	79.8	$20.00	1M	4.0
in $ 10.00 · out $ 50.00 · speed 34 tok/s · sources ls / bl · model id claude-fable-5 · news ↗
	Claude Fable 5.1 · Anthropic mid	65.0	$20.00	1M	3.3
in $ 10.00 · out $ 50.00 · speed 75 tok/s · sources ls / bl · model id claude-fable-5-1 · news ↗
	GPT-4 Turbo · OpenAI budget	48.0	$15.00	128K	3.2
in $ 10.00 · out $ 30.00 · speed - tok/s · sources ls / bl · model id gpt-4-turbo-2024-04-09 · no recent coverage matched
	o1-preview · OpenAI mid	57.3	$26.25	200K	2.2
in $ 15.00 · out $ 60.00 · speed - tok/s · sources ls / bl · model id o1-preview · no recent coverage matched
	o1 · OpenAI budget	46.2	$26.25	200K	1.8
in $ 15.00 · out $ 60.00 · speed - tok/s · sources - / bl · model id o1 · no recent coverage matched
	Claude Mythos Preview · Anthropic frontier	84.4	$50.00	0	1.7
in $ 25.00 · out $ 125 · speed - tok/s · sources ls / bl · model id claude-mythos-preview · no recent coverage matched
	Claude 3 Opus · Anthropic mid	50.4	$30.00	200K	1.7
in $ 15.00 · out $ 75.00 · speed - tok/s · sources ls / bl · model id claude-3-opus-20240229 · no recent coverage matched
	o3-pro · OpenAI budget	47.3	$35.00	200K	1.4
in $ 20.00 · out $ 80.00 · speed - tok/s · sources ls / bl · model id o3-pro-2025-06-10 · no recent coverage matched
	Claude 4.1 Opus · Anthropic budget	42.0	$30.00	200K	1.4
in $ 15.00 · out $ 75.00 · speed - tok/s · sources - / bl · model id claude-4-1-opus · no recent coverage matched
	GPT-5.2 Pro · OpenAI mid	64.9	$57.75	400K	1.1
in $ 21.00 · out $ 168 · speed - tok/s · sources ls / bl · model id gpt-5.2-pro-2025-12-11 · no recent coverage matched
	GPT-5.4 Pro · OpenAI mid	60.0	$67.50	1.1M	0.9
in $ 30.00 · out $ 180 · speed - tok/s · sources - / bl · model id gpt-5-4-pro · no recent coverage matched
	GPT-5.5 Pro · OpenAI mid	57.2	$67.50	1M	0.8
in $ 30.00 · out $ 180 · speed - tok/s · sources ls / bl · model id gpt-5.5-pro · no recent coverage matched
	o1-pro · OpenAI frontier	79.0	$263	200K	0.3
in $ 150 · out $ 600 · speed - tok/s · sources ls / bl · model id o1-pro · no recent coverage matched
	Hy4 preview · Tencent frontier open weights	92.3	0	1M	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id hy4-preview · no recent coverage matched
	GPT-5 Medium · OpenAI frontier	88.1	-	128K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id gpt-5-medium-2025-08-07 · no recent coverage matched
	GPT-5.1 High · OpenAI frontier	88.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-5.1-high-2025-11-12 · no recent coverage matched
	ChatGPT-4o Latest · OpenAI frontier	84.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id chatgpt-4o-latest · no recent coverage matched
	GPT-5.1 Thinking · OpenAI frontier	82.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-5.1-thinking-2025-11-12 · no recent coverage matched
	GPT OSS 120B High · OpenAI frontier open weights	80.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-oss-120b-high · no recent coverage matched
	EXAONE 4.5 33B · LG AI Research frontier	80.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id exaone-4.5-33b · no recent coverage matched
	Seed 2.0 Lite · ByteDance frontier	79.3	-	256K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id seed-2.0-lite · no recent coverage matched
	MAI-Thinking-1 · Microsoft frontier	78.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mai-thinking-1 · no recent coverage matched
	MiMo-V2-Pro · Xiaomi frontier	78.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mimo-v2-pro · no recent coverage matched
	Claude Opus 4.1 · Anthropic frontier	77.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-opus-4-1-20250805 · no recent coverage matched
	Claude 3.7 Sonnet · Anthropic frontier	77.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-3-7-sonnet-20250219 · no recent coverage matched
	Qwen3-Next-80B-A3B-Thinking · Alibaba Cloud / Qwen Team frontier open weights	77.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-next-80b-a3b-thinking · no recent coverage matched
	Qwen3.5-4B · Alibaba Cloud / Qwen Team frontier open weights	76.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3.5-4b · no recent coverage matched
	Claude Opus 4 · Anthropic frontier	76.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-opus-4-20250514 · no recent coverage matched
	Llama 3.1 Nemotron Ultra 253B v1 · NVIDIA frontier open weights	76.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id llama-3.1-nemotron-ultra-253b-v1 · no recent coverage matched
	Kimi K2 0905 · Moonshot AI frontier	75.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id kimi-k2-0905 · no recent coverage matched
	Nova 2 Pro · Amazon frontier	75.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id nova-2-pro · no recent coverage matched
	Muse Spark · Meta frontier	75.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id muse-spark · news ↗
	MiMo-V2-Omni · Xiaomi mid	74.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mimo-v2-omni · no recent coverage matched
	GPT-5 Codex · OpenAI mid	74.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-5-codex-2025-09-15 · no recent coverage matched
	Qwen3 VL 30B A3B Thinking · Alibaba Cloud / Qwen Team mid open weights	74.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-vl-30b-a3b-thinking · no recent coverage matched
	Gemini 2.0 Flash Thinking · Google mid	74.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemini-2.0-flash-thinking · no recent coverage matched
	GPT OSS 20B High · OpenAI mid open weights	74.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-oss-20b-high · no recent coverage matched
	Claude Sonnet 4 · Anthropic mid	74.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-sonnet-4-20250514 · no recent coverage matched
	ERNIE 4.5 · Baidu mid	74.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id ernie-4.5 · no recent coverage matched
	DeepSeek R1 Zero · DeepSeek mid open weights	73.3	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-r1-zero · no recent coverage matched
	Qwen3 VL 32B Thinking · Alibaba Cloud / Qwen Team mid open weights	73.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-vl-32b-thinking · no recent coverage matched
	Ministral 3 (14B Reasoning 2512) · Mistral AI mid open weights	71.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id ministral-14b-latest · no recent coverage matched
	LongCat-Flash-Thinking · Meituan mid open weights	70.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id longcat-flash-thinking · no recent coverage matched
	Qwen3 VL 8B Thinking · Alibaba Cloud / Qwen Team mid open weights	69.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-vl-8b-thinking · no recent coverage matched
	Grok-4 Heavy · xAI mid	69.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id grok-4-heavy · no recent coverage matched
	Kimi K2-Thinking-0905 · Moonshot AI mid open weights	68.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id kimi-k2-thinking-0905 · no recent coverage matched
	Phi 4 Reasoning Plus · Microsoft mid open weights	68.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id phi-4-reasoning-plus · no recent coverage matched
	Qwen3 VL 32B Instruct · Alibaba Cloud / Qwen Team mid open weights	68.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-vl-32b-instruct · no recent coverage matched
	Magistral Small 2506 · Mistral AI mid open weights	68.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id magistral-small-2506 · no recent coverage matched
	North Mini Code 1.0 · Cohere mid open weights	67.6	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id north-mini-code-1.0 · no recent coverage matched
	Qwen3.7-Plus · Alibaba Cloud / Qwen Team mid	67.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3.7-plus · no recent coverage matched
	LongCat-Flash-Chat · Meituan mid open weights	66.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id longcat-flash-chat · no recent coverage matched
	Ministral 3 (8B Reasoning 2512) · Mistral AI mid open weights	66.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id ministral-8b-latest · no recent coverage matched
	Llama-3.3 Nemotron Super 49B v1 · NVIDIA mid open weights	66.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id llama-3.3-nemotron-super-49b-v1 · no recent coverage matched
	Phi 4 Reasoning · Microsoft mid open weights	65.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id phi-4-reasoning · no recent coverage matched
	Ornith-1.5-397B · Ornith AI mid open weights	65.4	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ornith-1-5-397b · no recent coverage matched
	DeepSeek R1 Distill Llama 70B · DeepSeek mid open weights	65.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-r1-distill-llama-70b · no recent coverage matched
	QwQ-32B · Alibaba Cloud / Qwen Team mid open weights	65.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwq-32b · no recent coverage matched
	QwQ-32B-Preview · Alibaba Cloud / Qwen Team mid open weights	65.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwq-32b-preview · no recent coverage matched
	dots3-note Preview · Dots Studio mid open weights	64.4	0	512K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id dots3-note-preview · no recent coverage matched
	Qwen3 VL 4B Thinking · Alibaba Cloud / Qwen Team mid open weights	64.1	-	-	-
in $ - · out $ - · speed 22 tok/s · sources ls · model id qwen3-vl-4b-thinking · no recent coverage matched
	Nemotron Nano 9B v2 · NVIDIA mid open weights	64.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id nvidia-nemotron-nano-9b-v2 · no recent coverage matched
	Grok-4 · xAI mid	63.8	-	128K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id grok-4 · no recent coverage matched
	Qwen3.8-Flash-Next · Alibaba Cloud / Qwen Team mid open weights	63.8	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id qwen3.8-flash-next · news ↗
	DeepSeek R1 Distill Qwen 32B · DeepSeek mid open weights	62.1	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id deepseek-r1-distill-qwen-32b · no recent coverage matched
	Gemini 2.0 Flash · Google mid	62.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemini-2.0-flash · no recent coverage matched
	ERNIE 5.0 · Baidu mid	62.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id ernie-5.0 · no recent coverage matched
	GPT-5 · OpenAI mid	61.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-5-2025-08-07 · news ↗
	Devstral Medium · Mistral AI mid	61.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id devstral-medium-2507 · no recent coverage matched
	DeepSeek-V3.2 (Thinking) · DeepSeek mid open weights	60.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-reasoner · no recent coverage matched
	o1-mini · OpenAI mid	60.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id o1-mini · no recent coverage matched
	MiMo-V2-Flash · Xiaomi mid open weights	59.7	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id mimo-v2-flash · no recent coverage matched
	o1 · OpenAI mid	59.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id o1-2024-12-17 · no recent coverage matched
	Claude 3.5 Sonnet · Anthropic mid	59.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-3-5-sonnet-20240620 · no recent coverage matched
	DeepSeek R1 Distill Qwen 14B · DeepSeek mid open weights	59.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-r1-distill-qwen-14b · no recent coverage matched
	Gemini 3 Pro Deep Think · Google mid	58.9	-	2M	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id gemini-3-pro-deep-think · no recent coverage matched
	LongCat-Flash-Thinking-2601 · Meituan mid open weights	58.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id longcat-flash-thinking-2601 · no recent coverage matched
	Gemini 2.5 Pro Preview 06-05 · Google mid	58.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemini-2.5-pro-preview-06-05 · no recent coverage matched
	Claude 3.5 Sonnet · Anthropic mid	58.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-3-5-sonnet-20241022 · news ↗
	MAI-Code-1-Flash · Microsoft mid	58.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mai-code-1-flash · no recent coverage matched
	Grok-4.1 · xAI mid	57.6	-	1M	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id grok-4.1-2025-11-17 · no recent coverage matched
	Grok-2 · xAI mid	56.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id grok-2 · no recent coverage matched
	DeepSeek-V3.2-Exp · DeepSeek mid open weights	55.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-v3.2-exp · no recent coverage matched
	Qwen3 235B 2507 (Reasoning) · Alibaba mid open weights	55.7	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id qwen3-235b-2507-reasoning · no recent coverage matched
	Seed 2.1 Pro · ByteDance mid	55.7	-	256K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id seed-2.1-pro · no recent coverage matched
	Hy3 Preview · Tencent mid open weights	55.3	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id hy3-preview · no recent coverage matched
	GPT-5.3-Codex-Spark · OpenAI mid	54.6	-	256K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id gpt-5-3-codex-spark · no recent coverage matched
	Seed 2.1 Turbo · ByteDance mid	54.6	-	256K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id seed-2.1-turbo · no recent coverage matched
	Llama 3.1 Nemotron Nano 8B V1 · NVIDIA mid open weights	54.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id llama-3.1-nemotron-nano-8b-v1 · no recent coverage matched
	GPT-4.5 · OpenAI mid	53.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-4.5 · no recent coverage matched
	Qwen3 235B 2507 · Alibaba mid open weights	53.7	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id qwen3-235b-2507 · no recent coverage matched
	Devstral Small 1.1 · Mistral AI mid open weights	53.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id devstral-small-2507 · no recent coverage matched
	Min istral 3 (3B Reasoning 2512) · Mistral AI mid open weights	53.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id ministral-3b-latest · no recent coverage matched
	Grok 4 Fast · xAI mid	52.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id grok-4-fast · no recent coverage matched
	DeepSeek LLM 2.0 · DeepSeek mid open weights	52.3	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id deepseek-llm-2-0 · no recent coverage matched
	Phi 4 Mini Reasoning · Microsoft mid open weights	52.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id phi-4-mini-reasoning · no recent coverage matched
	DeepSeek-V3.2-Speciale · DeepSeek mid open weights	51.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-v3.2-speciale · no recent coverage matched
	Apodex 1.1 · Apodex mid	51.6	-	0	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id apodex-1-1 · no recent coverage matched
	Qwen3.5-2B · Alibaba Cloud / Qwen Team mid open weights	51.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3.5-2b · no recent coverage matched
	Gemini 2.0 Flash-Lite · Google mid	51.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemini-2.0-flash-lite · no recent coverage matched
	Gemini 1.5 Flash · Google mid	51.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemini-1.5-flash · no recent coverage matched
	Grok-2 mini · xAI mid	51.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id grok-2-mini · no recent coverage matched
	Llama 3.1 405B Instruct · Meta mid open weights	50.7	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id llama-3.1-405b-instruct · no recent coverage matched
	Sarvam-30B · Sarvam AI mid open weights	50.3	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id sarvam-30b · no recent coverage matched
	Qwen3-235B-A22B-Thinking-2507 · Alibaba Cloud / Qwen Team budget open weights	49.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-235b-a22b-thinking-2507 · no recent coverage matched
	GLM-4.7-Flash · Zhipu AI budget open weights	49.6	0	200K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id glm-4.7-flash · no recent coverage matched
	Qwen2.5 32B Instruct · Alibaba Cloud / Qwen Team budget open weights	49.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen-2.5-32b-instruct · no recent coverage matched
	DeepSeek R1 Distill Qwen 7B · DeepSeek budget open weights	49.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-r1-distill-qwen-7b · no recent coverage matched
	DeepSeek R1 Distill Llama 8B · DeepSeek budget open weights	49.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-r1-distill-llama-8b · no recent coverage matched
	DeepSeek V3.1 (Reasoning) · DeepSeek budget open weights	48.8	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id deepseek-v3-1-reasoning · no recent coverage matched
	Kimi K2-Instruct-0905 · Moonshot AI budget open weights	48.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id kimi-k2-instruct-0905 · no recent coverage matched
	DeepSeek Coder 2.0 · DeepSeek budget open weights	48.1	-	128K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id deepseek-coder-2-0 · no recent coverage matched
	Kimi K2 Base · Moonshot AI budget open weights	48.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id kimi-k2-base · no recent coverage matched
	Seed 1.6 · ByteDance budget	48.1	-	256K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id seed-1-6 · no recent coverage matched
	Nemotron 3 Super 100B · NVIDIA budget open weights	47.9	0	1M	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id nemotron-3-super-100b · no recent coverage matched
	o4-mini (high) · OpenAI budget	47.8	-	200K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id o4-mini-high · no recent coverage matched
	DeepSeekMath V2 · DeepSeek budget open weights	47.7	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id deepseekmath-v2 · no recent coverage matched
	Qwen2.5-1M · Alibaba budget open weights	47.7	0	1M	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id qwen2-5-1m · no recent coverage matched
	Qwen3 235B A22B · Alibaba Cloud / Qwen Team budget open weights	47.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-235b-a22b · no recent coverage matched
	Nova Pro · Amazon budget	46.9	-	128K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id nova-pro · no recent coverage matched
	Llama 3.2 90B Instruct · Meta budget open weights	46.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id llama-3.2-90b-instruct · no recent coverage matched
	Mistral Small 3.1 24B Instruct · Mistral AI budget open weights	46.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mistral-small-3.1-24b-instruct-2503 · no recent coverage matched
	Qwen2.5 VL 32B Instruct · Alibaba Cloud / Qwen Team budget open weights	46.0	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id qwen2.5-vl-32b · no recent coverage matched
	Qwen2.5 14B Instruct · Alibaba Cloud / Qwen Team budget open weights	45.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen-2.5-14b-instruct · no recent coverage matched
	Sarvam-105B · Sarvam AI budget open weights	45.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id sarvam-105b · no recent coverage matched
	MiniMax M1 80K · MiniMax budget open weights	44.8	-	80K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id minimax-m1-80k · no recent coverage matched
	MiniMax M1 40K · MiniMax budget open weights	44.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id minimax-m1-40k · no recent coverage matched
	Mistral Large 3 (675B Base) · Mistral AI budget open weights	43.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mistral-large-3-675b-base-2512 · no recent coverage matched
	Mistral Large 3 (675B Instruct 2512 Eagle) · Mistral AI budget open weights	43.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mistral-large-3-675B-instruct-2512-eagle · no recent coverage matched
	Mistral Large 3 (675B Instruct 2512 NVFP4) · Mistral AI budget open weights	43.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mistral-large-3-675b-instruct-2512-nvfp4 · no recent coverage matched
	Gemma 4 E2B · Google budget open weights	43.4	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id gemma-4-e2b-it · no recent coverage matched
	Z-1 · Z budget	43.1	-	128K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id z-1 · no recent coverage matched
	Seed 1.6 Flash · ByteDance budget	43.0	-	256K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id seed-1-6-flash · no recent coverage matched
	Mistral 8x7B · Mistral budget open weights	42.9	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id mistral-8x7b · no recent coverage matched
	Moonshot v1 · Moonshot AI budget	42.7	-	128K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id moonshot-v1 · no recent coverage matched
	DiffusionGemma 26B-A4B · Google budget open weights	42.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id diffusiongemma-26b-a4b-it · no recent coverage matched
	Qwen2 72B Instruct · Alibaba Cloud / Qwen Team budget open weights	42.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen2-72b-instruct · no recent coverage matched
	Nemotron Ultra 253B · NVIDIA budget open weights	42.3	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id nemotron-ultra-253b · no recent coverage matched
	Gemma 4 12B · Google budget open weights	42.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemma-4-12b-it · news ↗
	Nova Lite · Amazon budget	42.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id nova-lite · no recent coverage matched
	Claude 3.5 Haiku · Anthropic budget	41.1	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-3-5-haiku-20241022 · no recent coverage matched
	Nemotron 3 Nano Omni 30B A3B · NVIDIA budget open weights	40.6	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id nemotron-3-nano-omni-30b-a3b · no recent coverage matched
	Claude 3 Sonnet · Anthropic budget	40.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id claude-3-sonnet-20240229 · no recent coverage matched
	Ling 2.6 Flash · InclusionAI budget open weights	40.3	-	262K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id ling-2-6-flash · no recent coverage matched
	Nova Micro · Amazon budget	40.0	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id nova-micro · no recent coverage matched
	Kimi K2 Instruct · Moonshot AI budget open weights	39.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id kimi-k2-instruct · no recent coverage matched
	Magistral Medium · Mistral AI budget open weights	39.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id magistral-medium · no recent coverage matched
	Sarvam 105B · Sarvam budget open weights	39.8	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id sarvam-105b · no recent coverage matched
	LFM2.5-2.6B · Liquid AI budget open weights	39.6	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id lfm-2.5-2.6b · no recent coverage matched
	LFM2.5-8B-A1B · LiquidAI budget open weights	39.0	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-8b-a1b · no recent coverage matched
	Sarvam 30B · Sarvam budget open weights	38.6	0	64K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id sarvam-30b · no recent coverage matched
	Gemini 1.5 Flash 8B · Google budget	38.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemini-1.5-flash-8b · no recent coverage matched
	Qwen3-Omni-30B-A3B-Instruct · Alibaba budget open weights	38.3	-	0	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id qwen3-omni-30b-a3b-instruct · no recent coverage matched
	Granite-4.0-H-1B · IBM budget open weights	37.9	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id granite-4-0-h-1b · no recent coverage matched
	Granite-4.0-1B · IBM budget open weights	37.8	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id granite-4-0-1b · no recent coverage matched
	Llama 4 Behemoth · Meta budget open weights	37.8	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id llama-4-behemoth · no recent coverage matched
	Granite-4.0-350M · IBM budget open weights	37.7	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id granite-4-0-350m · no recent coverage matched
	Granite-4.0-H-350M · IBM budget open weights	37.7	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id granite-4-0-h-350m · no recent coverage matched
	Mistral Small 3.1 24B Base · Mistral AI budget open weights	37.5	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mistral-small-3.1-24b-base-2503 · no recent coverage matched
	Mistral 8x7B v0.2 · Mistral budget open weights	37.2	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id mistral-8x7b-v0-2 · no recent coverage matched
	Ornith-1.5-35B-A3B · Ornith AI budget open weights	37.1	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ornith-1-5-35b-a3b · no recent coverage matched
	Jamba 1.5 Large · AI21 Labs budget open weights	36.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id jamba-1.5-large · no recent coverage matched
	Phi-3.5-MoE-instruct · Microsoft budget open weights	36.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id phi-3.5-moe-instruct · no recent coverage matched
	Mistral Large 2 · Mistral budget	36.4	-	128K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id mistral-large-2 · no recent coverage matched
	Qwen2.5 7B Instruct · Alibaba Cloud / Qwen Team budget open weights	36.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen-2.5-7b-instruct · no recent coverage matched
	Grok-1.5 · xAI budget	35.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id grok-1.5 · no recent coverage matched
	GPT-4 · OpenAI budget	35.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gpt-4-0613 · news ↗
	Mistral Small 3 24B Base · Mistral AI budget open weights	34.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id mistral-small-24b-base-2501 · no recent coverage matched
	DeepSeek R1 Distill Qwen 1.5B · DeepSeek budget open weights	33.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-r1-distill-qwen-1.5b · no recent coverage matched
	Qwen2.5-Coder 32B Instruct · Alibaba Cloud / Qwen Team budget open weights	33.4	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id qwen-2.5-coder-32b-instruct · no recent coverage matched
	Llama 3.2 11B Instruct · Meta budget open weights	32.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id llama-3.2-11b-instruct · no recent coverage matched
	Llama 3.2 3B Instruct · Meta budget open weights	32.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id llama-3.2-3b-instruct · no recent coverage matched
	Jamba 1.5 Mini · AI21 Labs budget open weights	32.3	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id jamba-1.5-mini · no recent coverage matched
	Gemini Diffusion · Google budget	31.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemini-diffusion · no recent coverage matched
	Claude 4.1 Opus Thinking · Anthropic budget	31.6	-	200K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id claude-4-1-opus-thinking · no recent coverage matched
	Qwen2.5-Omni-7B · Alibaba Cloud / Qwen Team budget open weights	30.8	-	0	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id qwen2.5-omni-7b · no recent coverage matched
	Llama 3 70B · Meta budget open weights	30.6	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id llama-3-70b · no recent coverage matched
	Phi-3.5-mini-instruct · Microsoft budget open weights	30.4	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id phi-3.5-mini-instruct · no recent coverage matched
	Gemini 1.0 Pro · Google budget	27.9	-	32K	-
in $ - · out $ - · speed - tok/s · sources ls / bl · model id gemini-1.0-pro · no recent coverage matched
	Mixtral 8x22B Instruct v0.1 · Mistral budget open weights	26.7	0	64K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id mixtral-8x22b-instruct-v0-1 · no recent coverage matched
	Qwen2 7B Instruct · Alibaba Cloud / Qwen Team budget open weights	25.3	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen2-7b-instruct · no recent coverage matched
	Phi 4 Mini · Microsoft budget open weights	25.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id phi-4-mini · no recent coverage matched
	Gemma 3n E2B Instructed · Google budget	24.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemma-3n-e2b-it · no recent coverage matched
	Gemma 3n E2B Instructed LiteRT (Preview) · Google budget open weights	24.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemma-3n-e2b-it-litert-preview · no recent coverage matched
	Gemma 3n E4B Instructed · Google budget	23.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemma-3n-e4b-it · no recent coverage matched
	Gemma 3n E4B Instructed LiteRT Preview · Google budget open weights	23.7	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemma-3n-e4b-it-litert-preview · no recent coverage matched
	Gemma 3 1B · Google budget open weights	19.2	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id gemma-3-1b-it · no recent coverage matched
	LFM2-24B-A2B · LiquidAI budget	18.6	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-24b-a2b · no recent coverage matched
	DeepSeek-V2.5 · DeepSeek budget open weights	16.8	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id deepseek-v2.5 · no recent coverage matched
	LFM2.5-1.2B-Thinking · LiquidAI budget	16.7	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-1-2b-thinking · no recent coverage matched
	LFM2.5-1.2B-Instruct · LiquidAI budget	16.1	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-1-2b-instruct · no recent coverage matched
	Nemotron-4 15B · NVIDIA budget open weights	15.9	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id nemotron-4-15b · no recent coverage matched
	Qwen3 VL 235B A22B Thinking · Alibaba Cloud / Qwen Team budget open weights	13.6	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3-vl-235b-a22b-thinking · no recent coverage matched
	Qwen3.5-0.8B · Alibaba Cloud / Qwen Team budget open weights	11.9	-	-	-
in $ - · out $ - · speed - tok/s · sources ls · model id qwen3.5-0.8b · no recent coverage matched
	Mistral 7B v0.3 · Mistral budget open weights	8.7	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id mistral-7b-v0-3 · no recent coverage matched
	Laguna M.1 · Poolside budget	3.9	-	256K	-
in $ - · out $ - · speed - tok/s · sources - / bl · model id laguna-m-1 · no recent coverage matched
	Laguna XS.2 · Poolside budget open weights	1.5	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id laguna-xs-2 · no recent coverage matched
	1-bit Bonsai 1.7B · Prism ML open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id 1-bit-bonsai-1-7b · no recent coverage matched
	1-bit Bonsai 27B · Prism ML open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id 1-bit-bonsai-27b · no recent coverage matched
	1-bit Bonsai 4B · Prism ML open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id 1-bit-bonsai-4b · no recent coverage matched
	1-bit Bonsai 8B · Prism ML open weights	-	0	64K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id 1-bit-bonsai-8b · no recent coverage matched
	A.X K2 · SK Telecom open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id a-x-k2 · no recent coverage matched
	Agents-A1-4B · InternScience open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id agents-a1-4b · no recent coverage matched
	Apodex 1.1 Mini · Apodex open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id apodex-1-1-mini · no recent coverage matched
	BTL-3 · Bad Theory Labs open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id btl-3 · no recent coverage matched
	BTL-4 · Bad Theory Labs open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id btl-4 · no recent coverage matched
	BTL-4 Compact · Bad Theory Labs open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id btl-4-compact · no recent coverage matched
	Celeris-1 · Celeris	-	$0.325	128K	-
in $ 0.2 · out $ 0.7 · speed - tok/s · sources - / bl · model id celeris-1 · no recent coverage matched
	Claude Mythos 5 · Anthropic	-	$20.00	1M	-
in $ 10.00 · out $ 50.00 · speed - tok/s · sources - / bl · model id claude-mythos-5 · news ↗
	Composer 2 · Cursor	-	$1.00	200K	-
in $ 0.5 · out $ 2.50 · speed - tok/s · sources - / bl · model id composer-2 · no recent coverage matched
	Composer 2.5 · Cursor	-	$1.00	200K	-
in $ 0.5 · out $ 2.50 · speed - tok/s · sources - / bl · model id composer-2-5 · no recent coverage matched
	Cosmos3-Edge · NVIDIA open weights	-	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id cosmos3-edge · no recent coverage matched
	DBRX Instruct · Databricks open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id dbrx-instruct · no recent coverage matched
	DeepSeek-V3.2 (Non-thinking) · DeepSeek open weights	-	$0.315	131K	-
in $ 0.28 · out $ 0.42 · speed - tok/s · sources ls · model id deepseek-chat · no recent coverage matched
	DeepSeek-V4-Flash-0731 · DeepSeek open weights	-	$0.09	1.0M	-
in $ 0.06 · out $ 0.18 · speed - tok/s · sources ls / bl · model id deepseek-v4-flash-0731 · news ↗
	DeepSeek-V4-Flash-Vision-Exp · DeepSeek	-	$0.66	1.0M	-
in $ 0.44 · out $ 1.32 · speed - tok/s · sources ls · model id deepseek-v4-flash-vision-exp · no recent coverage matched
	Fara1.5-27B · Microsoft open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id fara1-5-27b · no recent coverage matched
	Fara1.5-4B · Microsoft open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id fara1-5-4b · no recent coverage matched
	Fugu Cyber · Sakana AI	-	$13.50	1M	-
in $ 6.00 · out $ 36.00 · speed - tok/s · sources - / bl · model id fugu-cyber · no recent coverage matched
	GPT Audio 1.5 · OpenAI	-	$4.38	128K	-
in $ 2.50 · out $ 10.00 · speed - tok/s · sources - / bl · model id gpt-audio-1-5 · no recent coverage matched
	GPT Realtime · OpenAI	-	$7.00	32K	-
in $ 4.00 · out $ 16.00 · speed - tok/s · sources - / bl · model id gpt-realtime · news ↗
	GPT Realtime 1.5 · OpenAI	-	$7.00	32K	-
in $ 4.00 · out $ 16.00 · speed - tok/s · sources - / bl · model id gpt-realtime-1-5 · no recent coverage matched
	GPT Realtime 2 · OpenAI	-	$9.00	128K	-
in $ 4.00 · out $ 24.00 · speed - tok/s · sources - / bl · model id gpt-realtime-2 · no recent coverage matched
	GPT Realtime 2.1 · OpenAI	-	$9.00	128K	-
in $ 4.00 · out $ 24.00 · speed - tok/s · sources - / bl · model id gpt-realtime-2-1 · no recent coverage matched
	GPT Realtime 2.1 Mini · OpenAI	-	$1.05	128K	-
in $ 0.6 · out $ 2.40 · speed - tok/s · sources - / bl · model id gpt-realtime-2-1-mini · no recent coverage matched
	GPT Realtime mini · OpenAI	-	$1.05	32K	-
in $ 0.6 · out $ 2.40 · speed - tok/s · sources - / bl · model id gpt-realtime-mini · no recent coverage matched
	GPT-4o Audio · OpenAI	-	$4.38	128K	-
in $ 2.50 · out $ 10.00 · speed - tok/s · sources - / bl · model id gpt-4o-audio · no recent coverage matched
	GPT-4o mini Audio · OpenAI	-	$0.263	128K	-
in $ 0.15 · out $ 0.6 · speed - tok/s · sources - / bl · model id gpt-4o-mini-audio · no recent coverage matched
	GPT-4o mini TTS · OpenAI	-	$3.45	2K	-
in $ 0.6 · out $ 12.00 · speed - tok/s · sources - / bl · model id gpt-4o-mini-tts · no recent coverage matched
	GPT-5.1 Medium · OpenAI	-	$3.44	400K	-
in $ 1.25 · out $ 10.00 · speed 61 tok/s · sources ls · model id gpt-5.1-medium-2025-11-12 · no recent coverage matched
	GPT-5.3 Chat · OpenAI	-	$4.81	128K	-
in $ 1.75 · out $ 14.00 · speed - tok/s · sources ls · model id gpt-5.3-chat-latest · no recent coverage matched
	GPT-5.6 Cyber · OpenAI	-	$28.13	400K	-
in $ 12.50 · out $ 75.00 · speed - tok/s · sources ls / bl · model id gpt-5.6-cyber · news ↗
	Grok Build 0.1 · xAI	-	$1.25	256K	-
in $ 1.00 · out $ 2.00 · speed - tok/s · sources - / bl · model id grok-build-0-1 · no recent coverage matched
	Grok-4 Fast Reasoning · xAI	-	$0.275	2M	-
in $ 0.2 · out $ 0.5 · speed 75 tok/s · sources ls · model id grok-4-fast-reasoning · no recent coverage matched
	Grok-4.1 Fast Non-Reasoning · xAI	-	$0.275	2M	-
in $ 0.2 · out $ 0.5 · speed 105 tok/s · sources ls · model id grok-4-1-fast-non-reasoning · no recent coverage matched
	Grok-4.1 Fast Reasoning · xAI	-	$0.275	2M	-
in $ 0.2 · out $ 0.5 · speed - tok/s · sources ls · model id grok-4-1-fast-reasoning · no recent coverage matched
	Hermes 3 405B · Nous Research open weights	-	$1.00	131K	-
in $ 1.00 · out $ 1.00 · speed - tok/s · sources ls · model id hermes-3-405b · no recent coverage matched
	Holo3-122B-A10B · H Company	-	$1.05	64K	-
in $ 0.4 · out $ 3.00 · speed - tok/s · sources - / bl · model id holo3-122b-a10b · no recent coverage matched
	Holo3.1-0.8B · H Company open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id holo3-1-0-8b · no recent coverage matched
	Holo3.1-35B-A3B · H Company open weights	-	$0.637	64K	-
in $ 0.25 · out $ 1.80 · speed - tok/s · sources - / bl · model id holo3-1-35b-a3b · no recent coverage matched
	Holo3.1-35B-A3B-FP8 · H Company open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id holo3-1-35b-a3b-fp8 · no recent coverage matched
	Holo3.1-35B-A3B-GGUF · H Company open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id holo3-1-35b-a3b-gguf · no recent coverage matched
	Holo3.1-35B-A3B-NVFP4 · H Company open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id holo3-1-35b-a3b-nvfp4 · no recent coverage matched
	Holo3.1-4B · H Company open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id holo3-1-4b · no recent coverage matched
	Holo3.1-9B · H Company open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id holo3-1-9b · no recent coverage matched
	Hy-MT2-1.8B · Tencent open weights	-	0	8K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id hy-mt2-1-8b · no recent coverage matched
	Hy-MT2-30B-A3B · Tencent open weights	-	0	8K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id hy-mt2-30b-a3b · no recent coverage matched
	Interfaze Beta · Interfaze	-	$2.00	1M	-
in $ 1.50 · out $ 3.50 · speed - tok/s · sources - / bl · model id interfaze-beta · no recent coverage matched
	K-EXAONE 2.0 · LG AI Research open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id k-exaone-2-0 · no recent coverage matched
	Kanana-2 1.3B Instruct · Kakao open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id kanana-2-1-3b-instruct · no recent coverage matched
	Kanana-2 3B Instruct · Kakao open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id kanana-2-3b-instruct · no recent coverage matched
	L3 8B Lunaris v1 · Sao10K	-	$0.043	8K	-
in $ 0.04 · out $ 0.05 · speed - tok/s · sources ls · model id l3-8b-lunaris-v1 · no recent coverage matched
	L3.1 70B Euryale v2.2 · Sao10K	-	$0.85	131K	-
in $ 0.85 · out $ 0.85 · speed - tok/s · sources ls · model id l3.1-70b-euryale-v2.2 · no recent coverage matched
	Laguna S 2.1 · Poolside open weights	-	$0.125	1.0M	-
in $ 0.1 · out $ 0.2 · speed - tok/s · sources ls / bl · model id laguna-s-2.1 · no recent coverage matched
	Leanstral · Mistral open weights	-	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id leanstral · no recent coverage matched
	Leanstral 1.5 · Mistral open weights	-	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id leanstral-1-5 · news ↗
	LFM2.5-230M · LiquidAI open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-230m · no recent coverage matched
	LFM2.5-350M · LiquidAI open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-350m · no recent coverage matched
	LFM2.5-ColBERT-350M · LiquidAI open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-colbert-350m · no recent coverage matched
	LFM2.5-Embedding-350M · LiquidAI open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-embedding-350m · no recent coverage matched
	LFM2.5-VL-3B · Liquid AI open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id lfm-2.5-vl-3b · no recent coverage matched
	LFM2.5-VL-450M · LiquidAI open weights	-	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id lfm2-5-vl-450m · no recent coverage matched
	Ling 3.0 Flash Fin · InclusionAI	-	$0.09	262K	-
in $ 0.06 · out $ 0.18 · speed - tok/s · sources ls / bl · model id ling-3.0-flash-fin · no recent coverage matched
	Ling 3.0 Flash VL · InclusionAI open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ling-3-0-flash-vl · no recent coverage matched
	Ling 3.0 Tiny · InclusionAI open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ling-3-0-tiny · no recent coverage matched
	LLaDA2.2-flash · InclusionAI open weights	-	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id llada2-2-flash · no recent coverage matched
	LLaDA2.2-mini · InclusionAI open weights	-	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id llada2-2-mini · no recent coverage matched
	Llama Guard 4 12B · Meta open weights	-	$0.18	164K	-
in $ 0.18 · out $ 0.18 · speed - tok/s · sources ls · model id llama-guard-4-12b · no recent coverage matched
	LongCat-2.0 · Meituan open weights	-	0	1M	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id longcat-2-0 · no recent coverage matched
	Macaw · Bad Theory Labs open weights	-	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id macaw · no recent coverage matched
	Mercury 2.5 · Inception	-	$0.068	260K	-
in $ 0.04 · out $ 0.15 · speed - tok/s · sources - / bl · model id mercury-2-5 · no recent coverage matched
	Mercury Edit 2 · Inception	-	$0.375	128K	-
in $ 0.25 · out $ 0.75 · speed - tok/s · sources - / bl · model id mercury-edit-2 · no recent coverage matched
	MiniCPM5-2B · OpenBMB open weights	-	0	131K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id minicpm5-2b · no recent coverage matched
	Mistral NeMo Instruct · Mistral AI open weights	-	$0.022	131K	-
in $ 0.019 · out $ 0.03 · speed - tok/s · sources ls · model id mistral-nemo-instruct-2407 · no recent coverage matched
	Mistral Small 4 (Reasoning) · Mistral open weights	-	$0.263	256K	-
in $ 0.15 · out $ 0.6 · speed - tok/s · sources - / bl · model id mistral-small-4-reasoning · no recent coverage matched
	Muse Spark 1.3 · Meta	-	$0.125	1.0M	-
in $ 0.1 · out $ 0.2 · speed 65 tok/s · sources ls / bl · model id muse-spark-1.3 · news ↗
	MythoMax L2 13B · Gryphe	-	$0.4	4K	-
in $ 0.4 · out $ 0.4 · speed - tok/s · sources ls · model id mythomax-l2-13b · no recent coverage matched
	Nemotron Content Safety 3.5 · NVIDIA open weights	-	$0.2	131K	-
in $ 0.2 · out $ 0.2 · speed - tok/s · sources ls · model id nemotron-content-safety-3.5 · no recent coverage matched
	North Micro Vision Instruct · Cohere open weights	-	0	128K	-
in $ 0 · out $ 0 · speed - tok/s · sources ls / bl · model id north-micro-vision-instruct · no recent coverage matched
	Nova 2 Sonic · Amazon	-	$0.935	1M	-
in $ 0.33 · out $ 2.75 · speed - tok/s · sources ls · model id nova-2-sonic · no recent coverage matched
	Ornith-1.0-35B · DeepReinforce AI open weights	-	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ornith-1-0-35b · no recent coverage matched
	Ornith-1.0-397B · DeepReinforce AI open weights	-	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ornith-1-0-397b · no recent coverage matched
	Ornith-1.0-9B · DeepReinforce AI open weights	-	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ornith-1-0-9b · no recent coverage matched
	Ornith-1.5-9B · Ornith AI open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ornith-1-5-9b · no recent coverage matched
	Pokee-Isaac 28B · Pokee AI	-	$0.362	10M	-
in $ 0.15 · out $ 1.00 · speed - tok/s · sources - / bl · model id pokee-isaac-28b · no recent coverage matched
	Quasar 438B · Multiverse Computing	-	$0.9	1M	-
in $ 0.6 · out $ 1.80 · speed - tok/s · sources - / bl · model id quasar-438b · no recent coverage matched
	Qwen-AgentWorld-35B-A3B · Alibaba open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id qwen-agentworld-35b-a3b · no recent coverage matched
	Qwen3 14B · Alibaba Cloud / Qwen Team open weights	-	$0.15	41K	-
in $ 0.12 · out $ 0.24 · speed - tok/s · sources ls · model id qwen3-14b · no recent coverage matched
	Qwen3 32B · Alibaba Cloud / Qwen Team open weights	-	$0.13	128K	-
in $ 0.08 · out $ 0.28 · speed - tok/s · sources ls · model id qwen3-32b · no recent coverage matched
	Qwen3 VL 235B A22B Instruct · Alibaba Cloud / Qwen Team open weights	-	$0.37	262K	-
in $ 0.2 · out $ 0.88 · speed - tok/s · sources ls · model id qwen3-vl-235b-a22b-instruct · no recent coverage matched
	Sakana Fugu-Ultra v1.1 · Sakana AI	-	$11.25	1M	-
in $ 5.00 · out $ 30.00 · speed - tok/s · sources - / bl · model id sakana-fugu-ultra-v1-1 · no recent coverage matched
	Sakana Namazu · Sakana AI	-	$1.71	256K	-
in $ 0.95 · out $ 4.00 · speed - tok/s · sources ls / bl · model id sakana-namazu · no recent coverage matched
	Seed 2.0 Code · ByteDance	-	$1.13	256K	-
in $ 0.5 · out $ 3.00 · speed - tok/s · sources ls · model id seed-2.0-code · no recent coverage matched
	Solar Open 2 · Upstage open weights	-	0	1M	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id solar-open-2 · no recent coverage matched
	Solar Pro 3 · Upstage	-	$0.263	128K	-
in $ 0.15 · out $ 0.6 · speed - tok/s · sources - / bl · model id solar-pro-3 · no recent coverage matched
	Soofi S 30B-A3B · Soofi Project open weights	-	0	1M	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id soofi-s-30b-a3b · no recent coverage matched
	Ternary Bonsai 1.7B · Prism ML open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ternary-bonsai-1-7b · no recent coverage matched
	Ternary Bonsai 27B · Prism ML open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ternary-bonsai-27b · no recent coverage matched
	Ternary Bonsai 4B · Prism ML open weights	-	0	32K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ternary-bonsai-4b · no recent coverage matched
	Ternary Bonsai 8B · Prism ML open weights	-	0	64K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ternary-bonsai-8b · no recent coverage matched
	Toast 1 · Mixedbread	-	$0.405	131K	-
in $ 0.3 · out $ 0.72 · speed - tok/s · sources - / bl · model id toast-1 · no recent coverage matched
	UI-Mate-27B · Tencent open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ui-mate-27b · no recent coverage matched
	UI-Mate-9B · Tencent open weights	-	0	262K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id ui-mate-9b · no recent coverage matched
	ZAYA1-74B-Preview · Zyphra open weights	-	0	256K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id zaya1-74b-preview · no recent coverage matched
	ZAYA1-8B · Zyphra open weights	-	0	131K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id zaya1-8b · no recent coverage matched
	ZAYA1-VL-8B · Zyphra open weights	-	0	131K	-
in $ 0 · out $ 0 · speed - tok/s · sources - / bl · model id zaya1-vl-8b · no recent coverage matched
no models match the filters

about the data

Where the numbers come from. Two public sources: benchlm.ai/llm-pricing (prices from September 11, 2026) and llm-stats.com, fetched 2026-09-15. Capability is the average of a model's available benchmark scores: GPQA, SWE-bench, HLE, each rescaled to 0-100. If a model has none, its benchlm score is used. Row tags: ls = llm-stats, bl = benchlm.
How we compute price. Blended $/1M = (3 x input + output) / 4. That is a 3:1 input-to-output token mix, typical for agent workloads. Hover any price for the raw input and output.
What the value score means. Capability divided by blended price: capability points per dollar. Higher means cheaper capability. It is not a quality ranking. Tiers by capability: frontier 75+, mid 50-74, budget under 50.
How the pareto frontier and quadrants are picked. A model is on the frontier when no other model is both at least as capable and at most as cheap. The quadrant split lines sit at the median blended price and median capability of the 50 plotted models; the frontier and quadrant presets use exactly the same lines. 8 models are on the frontier.
When it updates. Refreshed by a build from both sources, latest on 2026-09-15. Prices change without notice: verify with the provider before spending. Coverage: 492 rows shipped, 160 joined across both sources, 160 dropped for missing data. The same data ships as json at /api/llm/models.json.

News links are best effort: a row links to rotative coverage only when a headline clearly names that model.

← rotative