Skip to content
Bencher

Models

Every AI model we follow,Every model we track,side by side.

133 models from 24 companies. 117 of them are available on OpenRouter, a service that gives access to many models in one place. Click a model to see everything we know about it.133 models from 24 vendors, 117 of them routable on OpenRouter today. Sort by use-case composite, price or context, and open any model for every number we hold on it.

133 ofof 133
All tracked models
Gemini 4 Argon97.789.976.5Google (Gemini / DeepMind)30 Sep 2026——Mid-priced$2 / $10—
Claude Opus 5.586.681.289.9Anthropic22 Sep 2026—1MExpensive$4 / $20
Claude Fable 578.279.871.8Anthropic9 Jun 2026—1MExpensive$10 / $50
Claude Fable 5.174.980.181.8Anthropic1 Sep 2026—1MExpensive$10 / $50
Claude Opus 4.671.262.9—Anthropic4 Feb 2026—1MExpensive$5 / $25
Claude Opus 571.184.780.5Anthropic24 Jul 2026—1MExpensive$5 / $25
Claude Opus 4.769.5—41.0Anthropic16 Apr 2026—1MExpensive$5 / $25
Kimi K361.572.567.0Moonshot AI (Kimi)16 Jul 20261MExpensive$3 / $15
GPT-5.6 Sol61.371.158.3OpenAI9 Jul 2026—1.1MExpensive$4 / $20
Gemini 3 Pro60.0——Google (Gemini / DeepMind)18 Nov 2025————
Muse Spark 1.158.465.451.1Meta (Meta Superintelligence Labs, Muse)9 Jul 2026—1MMid-priced$1.25 / $4.25
GPT-6 Astra57.772.385.4OpenAI3 Sep 2026—1.1MExpensive$10 / $50
GLM-5.356.664.855.5Z.ai (Zhipu)14 Aug 20261MMid-priced$1.40 / $4.40
GPT-5.556.661.552.8OpenAI23 Apr 2026—1.1MExpensive$5 / $30
Gemini 3.8 Flash56.262.941.2Google (Gemini / DeepMind)2 Sep 2026—1MMid-priced$0.75 / $3.75
GPT-5.455.8—51.1OpenAI5 Mar 2026—1.1MMid-priced$2.50 / $15
Gemini 3.7 Flash55.871.0—Google (Gemini / DeepMind)13 Aug 2026—1MMid-priced$0.75 / $3.75
Muse Spark 1.255.568.6—Meta (Meta Superintelligence Labs, Muse)5 Aug 2026—1MMid-priced$1.25 / $4.25
Muse Spark 1.354.684.055.7Meta (Meta Superintelligence Labs, Muse)2 Sep 2026—1MMid-priced$1.25 / $4.25
Gemini 3.6 Flash53.4——Google (Gemini / DeepMind)21 Jul 2026—1MMid-priced$0.75 / $3.75
Claude Opus 4.852.657.152.3Anthropic28 May 2026—1MExpensive$5 / $25
Gemini 3.1 Pro (Preview)52.460.127.8Google (Gemini / DeepMind)19 Feb 2026—1MMid-priced$2 / $12
Claude Opus 4.550.3——Anthropic24 Nov 2025—200KExpensive$5 / $25
Qwen3.8 Max (0803)47.3——Qwen (Alibaba)—————
Gemini 3.5 Flash47.1——Google (Gemini / DeepMind)19 May 2026—1MMid-priced$1.50 / $9
MiMo-V2.6-Pro45.770.3—Xiaomi MiMo21 Sep 20261.1MCheap$0.43 / $0.87
DeepSeek V4.1 Flash38.552.4—DeepSeek10 Sep 20261MCheap$0.30 / $1.20
GPT-6 Sol36.549.764.0OpenAI22 Sep 2026—1.1MMid-priced$2 / $10
DeepSeek V4 Pro (0813)35.844.436.3DeepSeek13 Aug 20261MMid-priced$1.32 / $3.96
Qwen3.8 27B34.363.2—Qwen (Alibaba)14 Aug 20261MCheap$0.50 / $3
Claude Sonnet 532.655.836.0Anthropic30 Jun 2026—1MMid-priced$2 / $10
Grok 4.729.260.553.0SpaceXAI (formerly xAI)21 Sep 2026—500KMid-priced$2 / $6
GPT-6 Luna13.040.131.8OpenAI22 Sep 2026—1.1MCheap$0.10 / $0.50
Aion 3.5———AionLabs23 Sep 2026—262KMid-priced$3 / $6
Claude Sonnet 5.5—79.992.6Anthropic28 Sep 2026—1MMid-priced$2 / $10
Claude Mythos Preview———Anthropic7 Apr 2026————
Claude Sonnet 4.6———Anthropic17 Feb 2026—1MExpensive$3 / $15
Claude Haiku 4.5———Anthropic15 Oct 2025—200KMid-priced$1 / $5
Claude Sonnet 4.5———Anthropic29 Sep 2025—1MExpensive$3 / $15
SWE-1.7 Lightning———Cognition—————
SWE-2———Cognition—————
Command A+———Cohere20 May 2026128KCheap$0.30 / $1.50
Composer 2.5———Cursor—————
DeepSeek V4 Flash Vision Exp———DeepSeek21 Aug 20261MCheap$0.22 / $0.65
GPT-5.5 Pro———OpenAI23 Apr 2026—1.1MExpensive$30 / $180
GPT-5.4 Mini———OpenAI17 Mar 2026—400KMid-priced$0.75 / $4.50
DeepSeek V4 Flash (0731)———DeepSeek31 Jul 20261MCheap$0.02 / $1.28
DeepSeek V4 Pro (Preview, 0423)———DeepSeek24 Apr 20261MCheap$0.23 / $0.46
DeepSeek V4 Flash (Preview, 0423)———DeepSeek24 Apr 20261MCheap$0.04 / $0.08
DeepSeek V3.2———DeepSeek1 Dec 2025—164KCheap$0.28 / $0.42
Ember-1———Fireworks AI24 Sep 2026—1MExpensive$3 / $15
Gemini 3.5 Flash-Lite———Google (Gemini / DeepMind)21 Jul 2026—1MCheap$0.30 / $2.50
Gemma 4 12B IT———Google (Gemini / DeepMind)3 Jun 2026262K——
GPT-5.4 Pro———OpenAI5 Mar 2026—1.1MExpensive$30 / $180
GPT-5.3-Codex———OpenAI24 Feb 2026—400KMid-priced$1.75 / $14
Gemma 4 31B IT———Google (Gemini / DeepMind)2 Apr 2026262KCheap$0.09 / $0.34
Gemma 4 26B A4B IT———Google (Gemini / DeepMind)2 Apr 2026262KCheap$0.08 / $0.26
Gemini 3.1 Flash-Lite———Google (Gemini / DeepMind)3 Mar 2026—1MCheap$0.25 / $1.50
Gemini 3.1 Pro Preview Custom Tools———Google (Gemini / DeepMind)25 Feb 2026—1MMid-priced$2 / $12
Gemini 3 Flash Preview———Google (Gemini / DeepMind)17 Dec 2025—1MCheap$0.50 / $3
Granite 4.2 30B———IBM Granite25 Aug 2026131K——
Granite 4.2 8B———IBM Granite25 Aug 2026131KCheap$0.06 / $0.25
GPT-5.2-Codex———OpenAI14 Jan 2026—400KMid-priced$1.75 / $14
GPT-5.2———OpenAI10 Dec 2025—400KMid-priced$1.75 / $14
GPT-5.2 Pro———OpenAI10 Dec 2025—400KExpensive$21 / $168
Muse Glimmer 30B—26.2—Meta (Meta Superintelligence Labs, Muse)10 Aug 2026131KCheap$0.35 / $1.50
Muse Spark———Meta (Meta Superintelligence Labs, Muse)8 Apr 2026————
MiniMax M3—62.3—MiniMax1 Jun 20261MCheap$0.30 / $1.20
MiniMax M2.7———MiniMax18 Mar 2026205KCheap$0.30 / $1.20
MiniMax M2.5———MiniMax12 Feb 2026—205KCheap$0.27 / $1.08
MiniMax M2.1———MiniMax23 Dec 2025—205KCheap$0.30 / $1.20
Mistral Medium 3.5—21.1—Mistral AI28 Apr 2026262KMid-priced$1.50 / $7.50
GPT-5.2 Chat (ChatGPT snapshot)———OpenAI10 Dec 2025—128KMid-priced$1.75 / $14
GPT-5.1-Codex-Max———OpenAI4 Dec 2025—400KMid-priced$1.25 / $10
Mistral Small 4———Mistral AI16 Mar 2026262KCheap$0.15 / $0.60
Devstral 2———Mistral AI9 Dec 2025262KCheap$0.40 / $2
Ministral 3 14B———Mistral AI2 Dec 2025262KCheap$0.20 / $0.20
Kimi K2.7 Code———Moonshot AI (Kimi)12 Jun 2026262KMid-priced$0.95 / $4
Kimi K2.6———Moonshot AI (Kimi)20 Apr 2026262KMid-priced$0.95 / $4
Kimi K2.5———Moonshot AI (Kimi)27 Jan 2026—262KCheap$0.45 / $2.25
Nemotron 3.5 Lightning 30B-A3B———NVIDIA11 Aug 2026262KCheap$0.06 / $0.17
GPT-5.1———OpenAI13 Nov 2025—400KMid-priced$1.25 / $10
GPT-5 Mini———OpenAI7 Aug 2025—400KCheap$0.25 / $2
Nemotron 3 Ultra—47.1—NVIDIA4 Jun 2026262KCheap$0.60 / $2.40
GPT-5.6 Luna——41.2OpenAI9 Jul 2026—1.1MCheap$0.20 / $1.20
GPT-5.6 Luna Pro———OpenAI9 Jul 2026—1.1MCheap$0.20 / $1.20
Nemotron 3 Super 120B A12B———NVIDIA11 Mar 2026262KCheap$0.08 / $0.45
GPT-6.1 Sol—73.883.7OpenAI29 Sep 2026—1.1MMid-priced$2 / $10
GPT-6.1 Sol Pro———OpenAI29 Sep 2026—1.1MMid-priced$2 / $10
GPT-6 Sol Pro———OpenAI22 Sep 2026—1.1MMid-priced$2 / $10
GPT-6 Luna Pro———OpenAI22 Sep 2026—1.1MCheap$0.10 / $0.50
GPT-6 Astra Pro———OpenAI3 Sep 2026—1.1MExpensive$10 / $50
GPT-5.6 Sol Pro———OpenAI9 Jul 2026—1.1MExpensive$4 / $20
GPT-5.6 Terra—58.753.2OpenAI9 Jul 2026—1.1MMid-priced$2 / $12
GPT-5.6 Terra Pro———OpenAI9 Jul 2026—1.1MMid-priced$2 / $12
gpt-oss-120b———OpenAI5 Aug 2025131KCheap$0.04 / $0.17
gpt-oss-20b———OpenAI5 Aug 2025131KCheap$0.02 / $0.09
GPT-5.5 Codex———OpenAI—————
Qwen3.8 Max Prime———Qwen (Alibaba)22 Sep 2026—1MExpensive$4 / $12
Qwen3.8 Max (0902)—75.3—Qwen (Alibaba)2 Sep 2026—1MMid-priced$2 / $6
Qwen3.8 Flash———Qwen (Alibaba)26 Aug 20261MCheap$0.15 / $0.47
Qwen3.8 2.4T A95B (open weights)———Qwen (Alibaba)12 Aug 2026262KMid-priced$2 / $6
Qwen3.8 Max (0803)———Qwen (Alibaba)3 Aug 20261MMid-priced$2 / $6—
Qwen3.7 Plus———Qwen (Alibaba)21 May 2026—1MCheap$0.40 / $1.60
Qwen3.7 Max———Qwen (Alibaba)16 May 2026—1MMid-priced$2.50 / $7.50
Qwen3.6 35B A3B———Qwen (Alibaba)27 Apr 2026262KCheap$0.15 / $1
Qwen3.6 27B———Qwen (Alibaba)22 Apr 2026262KCheap$0.32 / $3.20
Qwen3.6 Max Preview———Qwen (Alibaba)18 Apr 2026—262KMid-priced$1.03 / $6.16
Qwen3.6 Plus———Qwen (Alibaba)2 Apr 2026—1MCheap$0.33 / $1.95
Qwen3.5-35B-A3B———Qwen (Alibaba)25 Feb 2026—262KCheap$0.16 / $1.30
Qwen3.8-Flash-Next———Qwen (Alibaba)————
Step 3.7 Flash———StepFun28 May 2026262KCheap$0.20 / $1.15
Step 5 Preview—76.0—StepFun—————
Apertus 1.5 70B———Swiss AI Initiative (EPFL, ETH Zurich, CSCS)24 Jul 2026262K——
Hy4 preview———Tencent Hunyuan28 Aug 20261MCheap$0.83 / $2.50
Hy3———Tencent Hunyuan6 Jul 2026262KCheap$0.13 / $0.53
Inkling Small———Thinking Machines Lab30 Jul 2026—524KCheap$0.45 / $1.20
Inkling—42.6—Thinking Machines Lab17 Jul 2026—524KMid-priced$1 / $4.05
EuroLLM 22B Instruct (2512)———EuroLLM (UTTER project consortium)5 Dec 202533K——
Grok 4.6——48.9SpaceXAI (formerly xAI)12 Aug 2026—500KMid-priced$2 / $6
Grok 4.5——41.4SpaceXAI (formerly xAI)8 Jul 2026—500KMid-priced$2 / $6
Grok Build 0.1———SpaceXAI (formerly xAI)19 May 2026—256KCheap$1 / $2
Grok 4.3———SpaceXAI (formerly xAI)30 Apr 2026—1MMid-priced$1.25 / $2.50
Grok 4.20 Multi-Agent———SpaceXAI (formerly xAI)31 Mar 2026—2MMid-priced$1.25 / $2.50
Grok 4.20———SpaceXAI (formerly xAI)10 Mar 2026—1MMid-priced$1.25 / $2.50
MiMo-V2.6-Flash———Xiaomi MiMo21 Sep 20261.1MCheap$0.14 / $0.28
MiMo-V2.5-Pro———Xiaomi MiMo22 Apr 2026—1.1MCheap$0.43 / $0.87
GLM-5.3 Prime———Z.ai (Zhipu)23 Sep 2026—1MMid-priced$2.80 / $8.80
GLM-5.3-FlashX———Z.ai (Zhipu)18 Sep 2026—1MCheap$0.37 / $1.25
GLM-5.3-Flash—74.6—Z.ai (Zhipu)26 Aug 20261MCheap$0.15 / $0.50
GLM-5.2——49.8Z.ai (Zhipu)16 Jun 20261MMid-priced$1.40 / $4.40
GLM-5.1———Z.ai (Zhipu)7 Apr 2026205KMid-priced$1.40 / $4.40
GLM 5———Z.ai (Zhipu)11 Feb 2026—205KCheap$0.60 / $1.92

Scores are out of 100 for each area (writing, research and analysis, coding). A dash means we don't have enough test results yet. Hover a score to see the model's rank and the thinking level it did best at.

Per-use-case composites (one row per model, best result per benchmark at any setting). Hover for rank and dominant setting. Prices are list $/1M tokens.