← Global
terminal
terminal ranking
index v0.2.0 · run cmsg3a05 · 2026-08-05 12:54 UTC| # | Model | Provider | Score (95% CI) | Consist. | Calib. | Contam. Δ | p50 | $/1k | |
|---|---|---|---|---|---|---|---|---|---|
| 1 |  OpenAI: GPT-5.6 Sol Pro | openai | 910 [822–997] | 1.00 | 1.00 | 0.000 | 420ms | $26.59 | |
| 2 |  MoonshotAI: Kimi K2.7 Code | moonshotai | 910 [822–997] | 1.00 | 1.00 | 0.000 | 770ms | $2.50 | |
| 3 |  Google: Gemma 4 26B A4B | 910 [822–997] | 1.00 | 1.00 | 0.000 | 340ms | $0.189 | ||
| 4 | OpenAI: GPT-5.6 Terra Pro | openai | 910 [822–997] | 1.00 | 1.00 | 0.000 | 364ms | $5.49 | |
| 5 | Google: Gemini 3.1 Pro Preview | 910 [822–997] | 1.00 | 1.00 | 0.000 | 2.5s | $19.20 | ||
| 6 | Google: Gemini 3.6 Flash | 910 [822–997] | 1.00 | 1.00 | 0.000 | 1.7s | $12.04 | ||
| 7 | OpenAI: gpt-oss-20b | openai | 909 [822–997] | 1.00 | 1.00 | 0.000 | 258ms | $0.150 | |
| 8 | ByteDance Seed: Seed-2.0-Lite | bytedance-seed | 909 [822–997] | 1.00 | 1.00 | 0.000 | 1.0s | $2.45 | |
| 9 | ByteDance Seed: Seed-2.0-Mini | bytedance-seed | 909 [822–997] | 1.00 | 1.00 | 0.000 | 448ms | $1.10 | |
| 10 |  Anthropic: Claude Sonnet 4.6 | anthropic | 908 [820–995] | 1.00 | 0.98 | 0.000 | 1.2s | $5.85 | |
| 11 | Anthropic: Claude Opus 4.8 (Fast) | anthropic | 908 [820–996] | 1.00 | 0.98 | 0.000 | 882ms | $20.01 | |
| 12 | Anthropic: Claude Opus 4.6 | anthropic | 908 [820–996] | 1.00 | 0.99 | 0.000 | 2.0s | $13.05 | |
| 13 | OpenAI: GPT-5.6 Luna Pro | openai | 907 [819–995] | 0.98 | 1.00 | 0.000 | 397ms | $0.757 | |
| 14 | Anthropic: Claude Sonnet 5 | anthropic | 905 [817–993] | 1.00 | 0.96 | 0.000 | 1.5s | $6.74 | |
| 15 |  DeepSeek: DeepSeek V3.1 | deepseek | 905 [817–992] | 0.97 | 1.00 | 0.000 | 1.2s | $0.501 | |
| 16 | Arcee AI: Trinity Large Thinking | arcee-ai | 887 [777–997] | 1.00 | 1.00 | 0.000 | 386ms | $2.48 | |
| 17 | AionLabs: Aion-3.0 | aion-labs | 858 [791–926] | 1.00 | 0.95 | 0.059 | 531ms | $6.70 | |
| 18 | Nex AGI: Nex-N2-Pro | nex-agi | 856 [761–952] | 1.00 | 0.97 | 0.038 | 749ms | $0.642 | |
| 19 | OpenAI: GPT-5.6 Sol | openai | 856 [761–952] | 1.00 | 0.97 | 0.038 | 405ms | $5.43 | |
| 20 | Google: Gemini 3.5 Flash | 856 [761–952] | 1.00 | 0.97 | 0.038 | 1.8s | $14.52 | ||
| 21 | Google: Gemma 4 31B | 856 [761–952] | 1.00 | 0.97 | 0.038 | 339ms | $0.181 | ||
| 22 | Anthropic: Claude Opus 4.8 | anthropic | 855 [760–950] | 0.98 | 0.98 | 0.038 | 1.3s | $9.83 | |
| 23 | DeepSeek: DeepSeek V3.2 | deepseek | 853 [758–948] | 0.98 | 0.96 | 0.038 | 1.2s | $0.261 | |
| 24 | DeepSeek: DeepSeek V3.2 Exp | deepseek | 851 [756–947] | 0.97 | 0.97 | 0.038 | 1.4s | $0.224 | |
| 25 | OpenAI: GPT-5.6 Terra | openai | 851 [756–947] | 0.97 | 0.97 | 0.038 | 378ms | $1.15 | |
| 26 | DeepSeek: DeepSeek V4 Pro | deepseek | 850 [749–951] | 0.93 | 0.94 | 0.000 | 1.3s | $0.871 | |
| 27 | Anthropic: Claude Fable 5 | anthropic | 843 [738–948] | 1.00 | 0.99 | 0.000 | 3.5s | $15.48 | |
| 28 |  NVIDIA: Nemotron 3 Super | nvidia | 824 [717–932] | 0.94 | 0.92 | 0.000 | 807ms | $0.547 | |
| 29 | Anthropic: Claude Opus 4.7 | anthropic | 806 [705–907] | 1.00 | 0.97 | 0.077 | 1.4s | $8.48 | |
| 30 | OpenAI: GPT-5.6 Luna | openai | 800 [699–901] | 0.98 | 0.93 | 0.077 | 341ms | $0.161 | |
| 31 | OpenAI: gpt-oss-120b | openai | 790 [681–898] | 0.98 | 0.85 | 0.000 | 354ms | $0.154 | |
| 32 | Nex AGI: Nex-N2-Mini | nex-agi | 774 [666–883] | 0.88 | 0.83 | 0.000 | 411ms | $0.069 | |
| 33 | Google: Gemini 3.1 Flash Lite Preview | 769 [659–878] | 0.98 | 0.80 | 0.000 | 407ms | $0.498 | ||
| 34 | DeepSeek: DeepSeek V4 Flash 0731 | deepseek | 750 [641–860] | 1.00 | 0.77 | 0.000 | 528ms | $0.188 | |
| 35 | Google: Gemini 3.5 Flash Lite | 705 [595–814] | 0.93 | 0.73 | 0.019 | 634ms | $0.648 | ||
| 36 | OpenAI: GPT-5.4 | openai | 698 [590–805] | 1.00 | 0.88 | 0.154 | 344ms | $1.16 | |
| 37 | OpenAI: GPT-5.5 | openai | 697 [589–804] | 1.00 | 0.87 | 0.154 | 403ms | $10.68 | |
| 38 | Inception: Mercury 2 | inception | 694 [587–802] | 0.98 | 0.87 | 0.154 | 126ms | $0.712 | |
| 39 | WizardLM-2 8x22B | microsoft | 681 [574–789] | 0.90 | 0.87 | 0.154 | 612ms | $0.774 | |
| 40 | Google: Gemini 3.1 Flash Lite | 660 [551–768] | 0.97 | 0.70 | 0.058 | 592ms | $0.533 | ||
| 41 |  Meta: Llama 3.3 70B Instruct | meta-llama | 657 [549–765] | 0.81 | 0.67 | 0.000 | 170ms | $0.043 | |
| 42 | OpenAI: GPT-5.4 Mini | openai | 649 [543–756] | 0.88 | 0.68 | 0.000 | 436ms | $0.370 | |
| 43 | NVIDIA: Nemotron 3 Ultra | nvidia | 645 [536–753] | 0.98 | 0.83 | 0.192 | 507ms | $5.52 | |
| 44 |  MiniMax: MiniMax M2.1 | minimax | 640 [536–745] | 1.00 | 0.60 | 0.000 | 915ms | $2.99 | |
| 45 | Magnum v4 72B | anthracite-org | 637 [531–743] | 0.83 | 0.63 | 0.000 | 752ms | $0.967 | |
| 46 | Meta: Llama 4 Maverick | meta-llama | 629 [525–734] | 0.93 | 0.60 | 0.000 | 304ms | $0.354 | |
| 47 | Perceptron: Perceptron Mk1 | perceptron | 617 [515–719] | 0.98 | 0.57 | 0.000 | 236ms | $0.100 | |
| 48 | Perplexity: Sonar | perplexity | 613 [508–717] | 0.82 | 0.60 | 0.000 | 1.8s | $0.284 | |
| 49 | DeepSeek: DeepSeek V4 Flash 0423 | deepseek | 600 [498–702] | 0.88 | 0.57 | 0.000 | 1.6s | $0.196 | |
| 50 |  Mistral: Mistral Medium 3.5 | mistralai | 591 [489–693] | 0.82 | 0.57 | 0.000 | 423ms | $2.80 | |
| 51 |  Cohere: Command A | cohere | 569 [473–666] | 0.95 | 0.51 | 0.000 | 252ms | $1.07 | |
| 52 | AI21: Jamba Large 1.7 | ai21 | 545 [452–638] | 0.93 | 0.47 | 0.000 | 256ms | $2.68 | |
| 53 |  Amazon: Nova Premier 1.0 | amazon | 540 [434–646] | 0.85 | 0.65 | 0.135 | 327ms | $3.43 | |
| 54 | Venice: Uncensored | cognitivecomputations | 507 [421–593] | 0.93 | 0.41 | 0.000 | 517ms | $0.199 | |
| 55 | Microsoft: Phi 4 | microsoft | 487 [394–580] | 0.74 | 0.47 | 0.038 | 129ms | $0.037 | |
| 56 | Mistral: Mistral Large 3 2512 | mistralai | 479 [401–557] | 1.00 | 0.34 | 0.000 | 886ms | $0.189 | |
| 57 | Mistral Large 2407 | mistralai | 472 [394–550] | 0.95 | 0.34 | 0.000 | 902ms | $0.764 | |
| 58 | MiniMax: MiniMax M2.7 | minimax | 454 [380–529] | 1.00 | 0.29 | 0.000 | 857ms | $3.02 | |
| 59 | Cohere: Command R (08-2024) | cohere | 445 [372–519] | 0.90 | 0.30 | 0.000 | 204ms | $0.066 | |
| 60 | Mistral: Mistral Small 4 | mistralai | 438 [360–516] | 0.73 | 0.34 | 0.000 | 608ms | $0.064 | |
| 61 | Mistral: Ministral 3 8B 2512 | mistralai | 419 [350–489] | 0.84 | 0.27 | 0.000 | 307ms | $0.047 | |
| 62 | Meta: Llama 4 Scout | meta-llama | 410 [345–475] | 0.90 | 0.24 | 0.000 | 211ms | $0.043 | |
| 63 | Claude Opus 5 (Fast) | anthropic | 405 [349–460] | 1.00 | — | 0.000 | 1.0s | $8.24 | |
| 64 | Claude Opus 5 | anthropic | 405 [349–460] | 1.00 | — | 0.000 | 2.1s | $3.85 | |
| 65 | OpenAI: GPT-5.4 Nano | openai | 396 [332–461] | 0.64 | 0.49 | 0.000 | 464ms | $0.168 | |
| 66 | Amazon: Nova 2 Lite | amazon | 395 [335–455] | 0.91 | 0.20 | 0.000 | 371ms | $0.171 | |
| 67 | NVIDIA: Nemotron 3 Nano 30B A3B | nvidia | 391 [333–449] | 1.00 | 0.16 | 0.000 | 107ms | $0.490 | |
| 68 | Amazon: Nova Lite 1.0 | amazon | 380 [320–440] | 0.80 | 0.21 | 0.000 | 364ms | $0.027 | |
| 69 | Morph: Morph V3 Fast | morph | 372 [332–411] | 1.00 | — | 0.000 | 258ms | $0.410 | |
| 70 | Mistral: Ministral 3 14B 2512 | mistralai | 370 [315–425] | 0.85 | 0.17 | 0.000 | 329ms | $0.062 | |
| 71 | Nous: Hermes 4 405B | nousresearch | 340 [294–386] | 0.83 | 0.10 | 0.000 | 343ms | $0.380 | |
| 72 | AionLabs: Aion-RP 1.0 (8B) | aion-labs | 306 [282–330] | 0.77 | 0.10 | 0.000 | 871ms | $0.251 | |
| 73 | IBM: Granite 4.1 8B | ibm-granite | 301 [267–336] | 0.79 | 0.00 | 0.000 | 121ms | $0.017 | |
| 74 | Cohere: Command R+ (08-2024) | cohere | 299 [225–372] | 0.90 | 0.33 | 0.231 | 246ms | $1.08 | |
| 75 | Meta: Llama 3.2 3B Instruct | meta-llama | 288 [252–323] | 0.66 | 0.05 | 0.000 | 223ms | $0.020 | |
| 76 | ByteDance: UI-TARS 7B | bytedance | 287 [245–329] | 0.55 | 0.07 | 0.000 | 232ms | $0.034 | |
| 77 | Nous: Hermes 4 70B | nousresearch | 285 [243–327] | 0.54 | 0.07 | 0.000 | 259ms | $0.049 | |
| 78 | IBM: Granite 4.0 Micro | ibm-granite | 283 [249–318] | 0.65 | 0.02 | 0.000 | 350ms | $0.008 | |
| 79 | Meta: Llama 3.2 1B Instruct | meta-llama | 272 [237–306] | 0.58 | 0.01 | 0.000 | 233ms | $0.019 | |
| 80 | Morph: Morph V3 Large | morph | 216 [175–256] | — | 0.00 | 0.000 | 269ms | $0.381 |
- 910 [822–997]consist 1.00calib 1.00Δ 0.000420ms$26.59/1k
- #2MoonshotAI: Kimi K2.7 Codemoonshotai910 [822–997]consist 1.00calib 1.00Δ 0.000770ms$2.50/1k
- #3Google: Gemma 4 26B A4B google910 [822–997]consist 1.00calib 1.00Δ 0.000340ms$0.189/1k
- 910 [822–997]consist 1.00calib 1.00Δ 0.000364ms$5.49/1k
- 910 [822–997]consist 1.00calib 1.00Δ 0.0002.5s$19.20/1k
- 910 [822–997]consist 1.00calib 1.00Δ 0.0001.7s$12.04/1k
- #7OpenAI: gpt-oss-20bopenai909 [822–997]consist 1.00calib 1.00Δ 0.000258ms$0.150/1k
- #8ByteDance Seed: Seed-2.0-Litebytedance-seed909 [822–997]consist 1.00calib 1.00Δ 0.0001.0s$2.45/1k
- #9ByteDance Seed: Seed-2.0-Minibytedance-seed909 [822–997]consist 1.00calib 1.00Δ 0.000448ms$1.10/1k
- #10Anthropic: Claude Sonnet 4.6anthropic908 [820–995]consist 1.00calib 0.98Δ 0.0001.2s$5.85/1k
- #11Anthropic: Claude Opus 4.8 (Fast)anthropic908 [820–996]consist 1.00calib 0.98Δ 0.000882ms$20.01/1k
- #12Anthropic: Claude Opus 4.6anthropic908 [820–996]consist 1.00calib 0.99Δ 0.0002.0s$13.05/1k
- 907 [819–995]consist 0.98calib 1.00Δ 0.000397ms$0.757/1k
- #14Anthropic: Claude Sonnet 5anthropic905 [817–993]consist 1.00calib 0.96Δ 0.0001.5s$6.74/1k
- #15DeepSeek: DeepSeek V3.1deepseek905 [817–992]consist 0.97calib 1.00Δ 0.0001.2s$0.501/1k
- 887 [777–997]consist 1.00calib 1.00Δ 0.000386ms$2.48/1k
- #17AionLabs: Aion-3.0aion-labs858 [791–926]consist 1.00calib 0.95Δ 0.059531ms$6.70/1k
- #18Nex AGI: Nex-N2-Pronex-agi856 [761–952]consist 1.00calib 0.97Δ 0.038749ms$0.642/1k
- #19OpenAI: GPT-5.6 Solopenai856 [761–952]consist 1.00calib 0.97Δ 0.038405ms$5.43/1k
- 856 [761–952]consist 1.00calib 0.97Δ 0.0381.8s$14.52/1k
- #21Google: Gemma 4 31Bgoogle856 [761–952]consist 1.00calib 0.97Δ 0.038339ms$0.181/1k
- #22Anthropic: Claude Opus 4.8anthropic855 [760–950]consist 0.98calib 0.98Δ 0.0381.3s$9.83/1k
- #23DeepSeek: DeepSeek V3.2deepseek853 [758–948]consist 0.98calib 0.96Δ 0.0381.2s$0.261/1k
- #24DeepSeek: DeepSeek V3.2 Expdeepseek851 [756–947]consist 0.97calib 0.97Δ 0.0381.4s$0.224/1k
- #25OpenAI: GPT-5.6 Terraopenai851 [756–947]consist 0.97calib 0.97Δ 0.038378ms$1.15/1k
- #26DeepSeek: DeepSeek V4 Prodeepseek850 [749–951]consist 0.93calib 0.94Δ 0.0001.3s$0.871/1k
- #27Anthropic: Claude Fable 5anthropic843 [738–948]consist 1.00calib 0.99Δ 0.0003.5s$15.48/1k
- 824 [717–932]consist 0.94calib 0.92Δ 0.000807ms$0.547/1k
- #29Anthropic: Claude Opus 4.7anthropic806 [705–907]consist 1.00calib 0.97Δ 0.0771.4s$8.48/1k
- #30OpenAI: GPT-5.6 Lunaopenai800 [699–901]consist 0.98calib 0.93Δ 0.077341ms$0.161/1k
- #31OpenAI: gpt-oss-120bopenai790 [681–898]consist 0.98calib 0.85Δ 0.000354ms$0.154/1k
- #32Nex AGI: Nex-N2-Mininex-agi774 [666–883]consist 0.88calib 0.83Δ 0.000411ms$0.069/1k
- 769 [659–878]consist 0.98calib 0.80Δ 0.000407ms$0.498/1k
- 750 [641–860]consist 1.00calib 0.77Δ 0.000528ms$0.188/1k
- 705 [595–814]consist 0.93calib 0.73Δ 0.019634ms$0.648/1k
- #36OpenAI: GPT-5.4openai698 [590–805]consist 1.00calib 0.88Δ 0.154344ms$1.16/1k
- #37OpenAI: GPT-5.5openai697 [589–804]consist 1.00calib 0.87Δ 0.154403ms$10.68/1k
- #38Inception: Mercury 2inception694 [587–802]consist 0.98calib 0.87Δ 0.154126ms$0.712/1k
- #39WizardLM-2 8x22Bmicrosoft681 [574–789]consist 0.90calib 0.87Δ 0.154612ms$0.774/1k
- 660 [551–768]consist 0.97calib 0.70Δ 0.058592ms$0.533/1k
- #41Meta: Llama 3.3 70B Instructmeta-llama657 [549–765]consist 0.81calib 0.67Δ 0.000170ms$0.043/1k
- #42OpenAI: GPT-5.4 Miniopenai649 [543–756]consist 0.88calib 0.68Δ 0.000436ms$0.370/1k
- 645 [536–753]consist 0.98calib 0.83Δ 0.192507ms$5.52/1k
- #44MiniMax: MiniMax M2.1minimax640 [536–745]consist 1.00calib 0.60Δ 0.000915ms$2.99/1k
- #45Magnum v4 72Banthracite-org637 [531–743]consist 0.83calib 0.63Δ 0.000752ms$0.967/1k
- #46Meta: Llama 4 Maverickmeta-llama629 [525–734]consist 0.93calib 0.60Δ 0.000304ms$0.354/1k
- #47Perceptron: Perceptron Mk1perceptron617 [515–719]consist 0.98calib 0.57Δ 0.000236ms$0.100/1k
- #48Perplexity: Sonarperplexity613 [508–717]consist 0.82calib 0.60Δ 0.0001.8s$0.284/1k
- 600 [498–702]consist 0.88calib 0.57Δ 0.0001.6s$0.196/1k
- #50Mistral: Mistral Medium 3.5mistralai591 [489–693]consist 0.82calib 0.57Δ 0.000423ms$2.80/1k
- #51Cohere: Command Acohere569 [473–666]consist 0.95calib 0.51Δ 0.000252ms$1.07/1k
- 545 [452–638]consist 0.93calib 0.47Δ 0.000256ms$2.68/1k
- 540 [434–646]consist 0.85calib 0.65Δ 0.135327ms$3.43/1k
- #54Venice: Uncensoredcognitivecomputations507 [421–593]consist 0.93calib 0.41Δ 0.000517ms$0.199/1k
- #55Microsoft: Phi 4microsoft487 [394–580]consist 0.74calib 0.47Δ 0.038129ms$0.037/1k
- #56Mistral: Mistral Large 3 2512mistralai479 [401–557]consist 1.00calib 0.34Δ 0.000886ms$0.189/1k
- #57Mistral Large 2407mistralai472 [394–550]consist 0.95calib 0.34Δ 0.000902ms$0.764/1k
- #58MiniMax: MiniMax M2.7minimax454 [380–529]consist 1.00calib 0.29Δ 0.000857ms$3.02/1k
- 445 [372–519]consist 0.90calib 0.30Δ 0.000204ms$0.066/1k
- #60Mistral: Mistral Small 4mistralai438 [360–516]consist 0.73calib 0.34Δ 0.000608ms$0.064/1k
- #61Mistral: Ministral 3 8B 2512mistralai419 [350–489]consist 0.84calib 0.27Δ 0.000307ms$0.047/1k
- #62Meta: Llama 4 Scoutmeta-llama410 [345–475]consist 0.90calib 0.24Δ 0.000211ms$0.043/1k
- #63Claude Opus 5 (Fast)anthropic405 [349–460]consist 1.00Δ 0.0001.0s$8.24/1k
- #64Claude Opus 5anthropic405 [349–460]consist 1.00Δ 0.0002.1s$3.85/1k
- #65OpenAI: GPT-5.4 Nanoopenai396 [332–461]consist 0.64calib 0.49Δ 0.000464ms$0.168/1k
- #66Amazon: Nova 2 Liteamazon395 [335–455]consist 0.91calib 0.20Δ 0.000371ms$0.171/1k
- 391 [333–449]consist 1.00calib 0.16Δ 0.000107ms$0.490/1k
- #68Amazon: Nova Lite 1.0amazon380 [320–440]consist 0.80calib 0.21Δ 0.000364ms$0.027/1k
- 372 [332–411]consist 1.00Δ 0.000258ms$0.410/1k
- #70Mistral: Ministral 3 14B 2512mistralai370 [315–425]consist 0.85calib 0.17Δ 0.000329ms$0.062/1k
- #71Nous: Hermes 4 405Bnousresearch340 [294–386]consist 0.83calib 0.10Δ 0.000343ms$0.380/1k
- #72AionLabs: Aion-RP 1.0 (8B)aion-labs306 [282–330]consist 0.77calib 0.10Δ 0.000871ms$0.251/1k
- #73IBM: Granite 4.1 8Bibm-granite301 [267–336]consist 0.79calib 0.00Δ 0.000121ms$0.017/1k
- 299 [225–372]consist 0.90calib 0.33Δ 0.231246ms$1.08/1k
- #75Meta: Llama 3.2 3B Instructmeta-llama288 [252–323]consist 0.66calib 0.05Δ 0.000223ms$0.020/1k
- #76ByteDance: UI-TARS 7B bytedance287 [245–329]consist 0.55calib 0.07Δ 0.000232ms$0.034/1k
- #77Nous: Hermes 4 70Bnousresearch285 [243–327]consist 0.54calib 0.07Δ 0.000259ms$0.049/1k
- #78IBM: Granite 4.0 Microibm-granite283 [249–318]consist 0.65calib 0.02Δ 0.000350ms$0.008/1k
- #79Meta: Llama 3.2 1B Instructmeta-llama272 [237–306]consist 0.58calib 0.01Δ 0.000233ms$0.019/1k
- 216 [175–256]calib 0.00Δ 0.000269ms$0.381/1k