Leaderboards

Benchmarks

Every score carries its origin.

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

Arena Creative Writing

Arena Score on creative-writing prompts — human preference, not a prose-craft rubric
Provenance68 of 168 modelsscores Sep 13, 2026 – Sep 25, 2026
101GPT-4.1 MiniOpenAIScore1349Score / $843ScoredSep 25, 2026102DeepSeek V3DeepSeek685BScore1349Score / $1515Score / GB VRAM3ScoredSep 25, 2026103Gemma 3 27BGoogle27.4BScore1348Score / $8427Score / GB VRAM~78ScoredSep 25, 2026104Step 3.5 FlashStepFun199BScore1345Score / $4483Score / GB VRAM~11ScoredSep 25, 2026105Qwen3.5-35B-A3BQwen36BScore1342Score / $1790Score / GB VRAM~59ScoredSep 25, 2026106Qwen3 VL 235B A22B ThinkingQwen236BScore1339Score / $339Score / GB VRAM~9ScoredSep 25, 2026107Qwen3.5-FlashQwenScore1338Score / $5146ScoredSep 25, 2026108GPT-4o (2024-05-13)OpenAIScore1338Score / $89ScoredSep 25, 2026109o4 MiniOpenAIScore1337Score / $304ScoredSep 25, 2026110GPT-5.4 NanoOpenAIScore1335Score / $2137ScoredSep 25, 2026111Gemma 3 12BGoogle12.2BScore1332Score / $13324Score / GB VRAM~170ScoredSep 25, 2026112Trinity Large ThinkingArcee AI399BScore1332Score / $1665Score / GB VRAM~5.3ScoredSep 25, 2026113GPT-4o (2024-08-06)OpenAIScore1327Score / $133ScoredSep 25, 2026114GLM 4.5 AirZ.ai111BScore1327Score / $1561Score / GB VRAM~19ScoredSep 25, 2026115Qwen3 Next 80B A3B ThinkingQwen81.3BScore1326Score / $1105Score / GB VRAM~26ScoredSep 25, 2026116GPT-5 MiniOpenAIScore1324Score / $1324ScoredSep 25, 2026117GPT-4 TurboOpenAIScore1323Score / $44ScoredSep 25, 2026118MiniMax M1MiniMax456BScore1319Score / $600Score / GB VRAM~4.6ScoredSep 25, 2026119Qwen3 30B A3B Instruct 2507Qwen30.5BScore1319Score / $6831Score / GB VRAM~69ScoredSep 25, 2026120Inkling SmallThinking Machines266BScore1313Score / $1094Score / GB VRAM~7.8ScoredSep 25, 2026121Qwen3 Next 80B A3B InstructQwen81.3BScore1313Score / $1683Score / GB VRAM~26ScoredSep 25, 2026122o3 Mini HighOpenAIScore1312Score / $298ScoredSep 25, 2026123GLM 4.5VZ.ai108BScore1309Score / $727Score / GB VRAM~19ScoredSep 25, 2026124Llama 4 MaverickMeta402BScore1307Score / $2003Score / GB VRAM~5.2ScoredSep 25, 2026125GPT-4.1 NanoOpenAIScore1307Score / $3267ScoredSep 25, 2026126GLM 4.7 FlashZ.ai31.2BScore1306Score / $3266Score / GB VRAM69ScoredSep 25, 2026127Qwen3 32BQwen32.8BScore1304Score / $4656Score / GB VRAM~63ScoredSep 25, 2026128Nemotron 3 SuperNVIDIA124BScore1303Score / $3259Score / GB VRAM~17ScoredSep 25, 2026129o3 MiniOpenAIScore1301Score / $296ScoredSep 25, 2026130Gemma 3n 4BGoogle7.8BScore1299Score / $10821Score / GB VRAM~260ScoredSep 25, 2026131GPT-4o-mini (2024-07-18)OpenAIScore1294Score / $2156ScoredSep 25, 2026132Gemma 2 27BGoogle27.2BScore1293Score / $1989Score / GB VRAM~75ScoredSep 25, 2026133Mercury 2InceptionScore1292Score / $1723ScoredSep 25, 2026134Llama 4 ScoutMeta109BScore1289Score / $4295Score / GB VRAM~19ScoredSep 25, 2026135Mistral Large 2407Mistral AIScore1287Score / $215ScoredSep 25, 2026136Llama 3.3 70B InstructMeta70.6BScore1285Score / $4017Score / GB VRAM~29ScoredSep 25, 2026137MiniMax M2MiniMax229BScore1285Score / $1260Score / GB VRAM~8.9ScoredSep 25, 2026138gpt-oss-120bOpenAI120BScore1277Score / $7514Score / GB VRAM~17ScoredSep 25, 2026139Nemotron 3.5 LightningNVIDIA31.6BScore1277Score / $7979Score / GB VRAM~64ScoredSep 25, 2026140Gemma 3 4BGoogle4.3BScore1274Score / $12745Score / GB VRAM~470ScoredSep 25, 2026141Mistral Small 3.1 24BMistral AI24BScore1270Score / $2289Score / GB VRAM~84ScoredSep 25, 2026142Nova 2 LiteAmazonScore1269Score / $508ScoredSep 25, 2026143GPT-4OpenAIScore1267Score / $21ScoredSep 25, 2026144Ling-2.6-flashinclusionAI107BScore1267Score / $42221Score / GB VRAM~19ScoredSep 13, 2026145Olmo 3 32B ThinkAllen Institute for AI32.2BScore1265Score / $2529Score / GB VRAM~62ScoredSep 25, 2026146Jamba Large 1.7AI21399BScore1261Score / $158Score / GB VRAM~5ScoredSep 13, 2026147Granite 4.1 8BIBM8.8BScore1260Score / $12601Score / GB VRAM~230ScoredSep 25, 2026148Llama 3.1 70B InstructMeta70.6BScore1257Score / $3142Score / GB VRAM~28ScoredSep 25, 2026149Qwen2.5 72B InstructQwen72.7BScore1254Score / $3136Score / GB VRAM~27ScoredSep 25, 2026150GPT-5 NanoOpenAIScore1247Score / $6233ScoredSep 25, 2026151Nemotron 3 Nano 30B A3BNVIDIA31.6BScore1241Score / $6207Score / GB VRAM~62ScoredSep 25, 2026152gpt-oss-safeguard-20bOpenAI21.5BScore1240Score / $13775Score / GB VRAM~91ScoredSep 25, 2026153Nova Pro 1.0AmazonScore1237Score / $386ScoredSep 25, 2026154Mistral Small 3Mistral AI23.6BScore1227Score / $15334Score / GB VRAM~82ScoredSep 25, 2026155Ministral 8BMistral AIScore1222Score / $11105ScoredSep 25, 2026156Nova Lite 1.0AmazonScore1221Score / $5085ScoredSep 25, 2026157Claude 3 HaikuAnthropicScore1216Score / $973ScoredSep 25, 2026158Mistral LargeMistral AIScore1211Score / $202ScoredSep 25, 2026159Phi 4Microsoft14.7BScore1210Score / $8644Score / GB VRAM134ScoredSep 25, 2026160Command R (08-2024)CohereScore1209Score / $2015ScoredSep 25, 2026161Qwen2.5 Coder 32B InstructQwen32.8BScore1207Score / $1207Score / GB VRAM~58ScoredSep 25, 2026162Granite 4.2 8BIBM8.8BScore1204Score / $8024Score / GB VRAM~220ScoredSep 25, 2026163Nova Micro 1.0AmazonScore1197Score / $8554ScoredSep 25, 2026164GPT-3.5 Turbo (older v0613)OpenAIScore1191Score / $595ScoredSep 25, 2026165Mixtral 8x22B InstructMistral AI141BScore1191Score / $198Score / GB VRAM~13ScoredSep 25, 2026166Llama 3.1 8B InstructMeta8BScore1177Score / $29426Score / GB VRAM~230ScoredSep 25, 2026167Llama 3.2 3B InstructMeta3.2BScore1145Score / $22892Score / GB VRAM~570ScoredSep 25, 2026168Llama 3.2 1B InstructMeta1.2BScore1083Score / $54127Score / GB VRAM~1400ScoredSep 25, 2026

Showing 101–168 of 168 models on this board.