Leaderboards

Benchmarks

Every score carries its origin.

Reasoning374Coding337Math221Agentic375Chat636Speech778
Arena Creative Writing168Arena Instruction Following168Arena Text (overall)168IFEval16LiveBench Instruction Following58LiveBench Language58

Arena Instruction Following

Arena Score on instruction-following prompts — not a percentage
Provenance68 of 168 modelsscores Sep 13, 2026 – Sep 25, 2026
101Qwen3 VL 235B A22B ThinkingQwen236BScore1383Score / $350Score / GB VRAM~9.3ScoredSep 25, 2026102Qwen3.5-FlashQwenScore1381Score / $5312ScoredSep 25, 2026103MiniMax M2.5MiniMax229BScore1381Score / $1454Score / GB VRAM~9.6ScoredSep 25, 2026104DeepSeek V3 0324DeepSeek685BScore1378Score / $1531Score / GB VRAM~3.2ScoredSep 25, 2026105Qwen3 Next 80B A3B InstructQwen81.3BScore1376Score / $1765Score / GB VRAM~27ScoredSep 25, 2026106Claude Sonnet 5AnthropicScore1374Score / $137ScoredSep 13, 2026107GPT-5 MiniOpenAIScore1373Score / $1373ScoredSep 25, 2026108GPT-4.1 MiniOpenAIScore1373Score / $858ScoredSep 25, 2026109o4 MiniOpenAIScore1369Score / $311ScoredSep 25, 2026110GLM 4.6VZ.ai108BScore1367Score / $1519Score / GB VRAM~20ScoredSep 25, 2026111Qwen3 30B A3B Instruct 2507Qwen30.5BScore1366Score / $7078Score / GB VRAM~71ScoredSep 25, 2026112o3 Mini HighOpenAIScore1366Score / $311ScoredSep 25, 2026113GLM 4.5 AirZ.ai111BScore1361Score / $1601Score / GB VRAM~20ScoredSep 25, 2026114Qwen3 Next 80B A3B ThinkingQwen81.3BScore1359Score / $1133Score / GB VRAM~27ScoredSep 25, 2026115Trinity Large ThinkingArcee AI399BScore1357Score / $1697Score / GB VRAM~5.4ScoredSep 25, 2026116GLM 4.7 FlashZ.ai31.2BScore1348Score / $3370Score / GB VRAM71ScoredSep 25, 2026117MiniMax M1MiniMax456BScore1346Score / $612Score / GB VRAM~4.7ScoredSep 25, 2026118o3 MiniOpenAIScore1344Score / $305ScoredSep 25, 2026119DeepSeek V3DeepSeek685BScore1343Score / $1509Score / GB VRAM3ScoredSep 25, 2026120Gemma 3 27BGoogle27.4BScore1343Score / $8395Score / GB VRAM~78ScoredSep 25, 2026121Nemotron 3.5 LightningNVIDIA31.6BScore1343Score / $8393Score / GB VRAM~67ScoredSep 25, 2026122Nemotron 3 SuperNVIDIA124BScore1343Score / $3356Score / GB VRAM~17ScoredSep 25, 2026123GLM 4.5VZ.ai108BScore1338Score / $744Score / GB VRAM~20ScoredSep 25, 2026124MiniMax M2MiniMax229BScore1334Score / $1308Score / GB VRAM~9.2ScoredSep 25, 2026125Qwen3 32BQwen32.8BScore1331Score / $4754Score / GB VRAM~64ScoredSep 25, 2026126Nova 2 LiteAmazonScore1327Score / $531ScoredSep 25, 2026127GPT-5 NanoOpenAIScore1327Score / $6635ScoredSep 25, 2026128Mercury 2InceptionScore1327Score / $1769ScoredSep 25, 2026129GPT-4o (2024-05-13)OpenAIScore1326Score / $88ScoredSep 25, 2026130gpt-oss-120bOpenAI120BScore1323Score / $7784Score / GB VRAM~17ScoredSep 25, 2026131Gemma 3 12BGoogle12.2BScore1321Score / $13211Score / GB VRAM~170ScoredSep 25, 2026132GPT-4o (2024-08-06)OpenAIScore1319Score / $132ScoredSep 25, 2026133Ling-2.6-flashinclusionAI107BScore1315Score / $43841Score / GB VRAM~19ScoredSep 13, 2026134Llama 4 MaverickMeta402BScore1314Score / $2013Score / GB VRAM~5.2ScoredSep 25, 2026135GPT-4 TurboOpenAIScore1303Score / $43ScoredSep 25, 2026136GPT-4.1 NanoOpenAIScore1300Score / $3251ScoredSep 25, 2026137Mistral Large 2407Mistral AIScore1299Score / $217ScoredSep 25, 2026138Llama 4 ScoutMeta109BScore1299Score / $4330Score / GB VRAM~19ScoredSep 25, 2026139Olmo 3 32B ThinkAllen Institute for AI32.2BScore1298Score / $2597Score / GB VRAM~64ScoredSep 25, 2026140Mistral Small 3.1 24BMistral AI24BScore1294Score / $2331Score / GB VRAM~86ScoredSep 25, 2026141GPT-4o-mini (2024-07-18)OpenAIScore1293Score / $2155ScoredSep 25, 2026142Llama 3.3 70B InstructMeta70.6BScore1293Score / $4039Score / GB VRAM~29ScoredSep 25, 2026143Qwen2.5 72B InstructQwen72.7BScore1292Score / $3230Score / GB VRAM~28ScoredSep 25, 2026144Nemotron 3 Nano 30B A3BNVIDIA31.6BScore1291Score / $6455Score / GB VRAM~65ScoredSep 25, 2026145Granite 4.1 8BIBM8.8BScore1287Score / $12874Score / GB VRAM~230ScoredSep 25, 2026146GPT-4OpenAIScore1285Score / $21ScoredSep 25, 2026147Gemma 3n 4BGoogle7.8BScore1282Score / $10679Score / GB VRAM~260ScoredSep 25, 2026148gpt-oss-safeguard-20bOpenAI21.5BScore1281Score / $14237Score / GB VRAM~94ScoredSep 25, 2026149Granite 4.2 8BIBM8.8BScore1280Score / $8531Score / GB VRAM~230ScoredSep 25, 2026150Nova Pro 1.0AmazonScore1277Score / $399ScoredSep 25, 2026151Llama 3.1 70B InstructMeta70.6BScore1273Score / $3182Score / GB VRAM~29ScoredSep 25, 2026152Gemma 2 27BGoogle27.2BScore1271Score / $1956Score / GB VRAM~74ScoredSep 25, 2026153Gemma 3 4BGoogle4.3BScore1267Score / $12675Score / GB VRAM~470ScoredSep 25, 2026154Jamba Large 1.7AI21399BScore1267Score / $158Score / GB VRAM~5ScoredSep 13, 2026155Qwen2.5 Coder 32B InstructQwen32.8BScore1264Score / $1264Score / GB VRAM~61ScoredSep 25, 2026156Mistral Small 3Mistral AI23.6BScore1255Score / $15692Score / GB VRAM~84ScoredSep 25, 2026157Phi 4Microsoft14.7BScore1245Score / $8896Score / GB VRAM138ScoredSep 25, 2026158Claude 3 HaikuAnthropicScore1245Score / $996ScoredSep 25, 2026159Nova Lite 1.0AmazonScore1244Score / $5182ScoredSep 25, 2026160Mistral LargeMistral AIScore1238Score / $206ScoredSep 25, 2026161Command R (08-2024)CohereScore1236Score / $2060ScoredSep 25, 2026162GPT-3.5 Turbo (older v0613)OpenAIScore1216Score / $608ScoredSep 25, 2026163Mixtral 8x22B InstructMistral AI141BScore1215Score / $203Score / GB VRAM~14ScoredSep 25, 2026164Nova Micro 1.0AmazonScore1215Score / $8678ScoredSep 25, 2026165Ministral 8BMistral AIScore1211Score / $11010ScoredSep 25, 2026166Llama 3.1 8B InstructMeta8BScore1191Score / $29773Score / GB VRAM~240ScoredSep 25, 2026167Llama 3.2 3B InstructMeta3.2BScore1146Score / $22926Score / GB VRAM~570ScoredSep 25, 2026168Llama 3.2 1B InstructMeta1.2BScore1086Score / $54295Score / GB VRAM~1400ScoredSep 25, 2026

Showing 101–168 of 168 models on this board.