Leaderboards

Benchmarks

Every score carries its origin.

Reasoning374Coding337Math221Agentic375Chat636Speech778
Arena Maths163LiveBench Mathematics58

Arena Maths

Arena Score on maths prompts — not a percentage
Provenance63 of 163 modelsscores Sep 13, 2026 – Sep 25, 2026
101DeepSeek V3.1 TerminusDeepSeek685BScore1397Score / $1397Score / GB VRAM~3.2ScoredSep 25, 2026102Qwen3 235B A22B Thinking 2507Qwen235BScore1397Score / $607Score / GB VRAM~9.4ScoredSep 25, 2026103MiniMax M2.5MiniMax229BScore1393Score / $1466Score / GB VRAM~9.7ScoredSep 25, 2026104Qwen3 Next 80B A3B ThinkingQwen81.3BScore1393Score / $1160Score / GB VRAM~27ScoredSep 25, 2026105Claude Sonnet 4AnthropicScore1390Score / $93ScoredSep 25, 2026106GLM 4.5 AirZ.ai111BScore1387Score / $1632Score / GB VRAM~20ScoredSep 25, 2026107Trinity Large ThinkingArcee AI399BScore1383Score / $1729Score / GB VRAM~5.5ScoredSep 25, 2026108o3 MiniOpenAIScore1381Score / $314ScoredSep 25, 2026109gpt-oss-120bOpenAI120BScore1381Score / $8122Score / GB VRAM~18ScoredSep 25, 2026110Qwen3 30B A3B Instruct 2507Qwen30.5BScore1378Score / $7138Score / GB VRAM~72ScoredSep 25, 2026111GPT-4.1OpenAIScore1375Score / $172ScoredSep 25, 2026112Qwen3 Coder 480B A35BQwen480BScore1374Score / $1374Score / GB VRAM~4.5ScoredSep 25, 2026113Nemotron 3 SuperNVIDIA124BScore1372Score / $3430Score / GB VRAM~18ScoredSep 25, 2026114MiniMax M1MiniMax456BScore1370Score / $623Score / GB VRAM~4.8ScoredSep 25, 2026115DeepSeek V3 0324DeepSeek685BScore1368Score / $1520Score / GB VRAM~3.2ScoredSep 25, 2026116Nemotron 3.5 LightningNVIDIA31.6BScore1365Score / $8531Score / GB VRAM~69ScoredSep 25, 2026117GLM 4.7 FlashZ.ai31.2BScore1361Score / $3404Score / GB VRAM72ScoredSep 25, 2026118GLM 4.5VZ.ai108BScore1360Score / $756Score / GB VRAM~20ScoredSep 25, 2026119GPT-4.1 MiniOpenAIScore1353Score / $846ScoredSep 25, 2026120Claude Sonnet 5AnthropicScore1351Score / $135ScoredSep 13, 2026121MiniMax M2MiniMax229BScore1351Score / $1325Score / GB VRAM~9.4ScoredSep 25, 2026122Ling-2.6-flashinclusionAI107BScore1349Score / $44959Score / GB VRAM~20ScoredSep 13, 2026123Nemotron 3 Nano 30B A3BNVIDIA31.6BScore1347Score / $6734Score / GB VRAM~68ScoredSep 25, 2026124GPT-5 NanoOpenAIScore1346Score / $6731ScoredSep 25, 2026125gpt-oss-safeguard-20bOpenAI21.5BScore1335Score / $14835Score / GB VRAM~98ScoredSep 25, 2026126Nova 2 LiteAmazonScore1332Score / $533ScoredSep 25, 2026127Gemma 3 27BGoogle27.4BScore1322Score / $8265Score / GB VRAM~77ScoredSep 25, 2026128Gemma 3 12BGoogle12.2BScore1318Score / $13177Score / GB VRAM~170ScoredSep 25, 2026129Granite 4.1 8BIBM8.8BScore1317Score / $13167Score / GB VRAM~240ScoredSep 25, 2026130Llama 4 MaverickMeta402BScore1316Score / $2017Score / GB VRAM~5.2ScoredSep 25, 2026131Olmo 3 32B ThinkAllen Institute for AI32.2BScore1311Score / $2621Score / GB VRAM~65ScoredSep 25, 2026132DeepSeek V3DeepSeek685BScore1310Score / $1472Score / GB VRAM3ScoredSep 25, 2026133Llama 4 ScoutMeta109BScore1309Score / $4362Score / GB VRAM~19ScoredSep 25, 2026134GPT-4o (2024-08-06)OpenAIScore1309Score / $131ScoredSep 25, 2026135GPT-4o (2024-05-13)OpenAIScore1305Score / $87ScoredSep 25, 2026136GPT-4 TurboOpenAIScore1296Score / $43ScoredSep 25, 2026137Qwen2.5 72B InstructQwen72.7BScore1296Score / $3240Score / GB VRAM~28ScoredSep 25, 2026138Llama 3.3 70B InstructMeta70.6BScore1296Score / $4050Score / GB VRAM~29ScoredSep 25, 2026139Mistral Large 2407Mistral AIScore1288Score / $215ScoredSep 25, 2026140GPT-4OpenAIScore1284Score / $21ScoredSep 25, 2026141Mistral Small 3.1 24BMistral AI24BScore1277Score / $2301Score / GB VRAM~84ScoredSep 25, 2026142GPT-4o-mini (2024-07-18)OpenAIScore1275Score / $2126ScoredSep 25, 2026143GPT-4.1 NanoOpenAIScore1274Score / $3184ScoredSep 25, 2026144Qwen2.5 Coder 32B InstructQwen32.8BScore1270Score / $1270Score / GB VRAM~61ScoredSep 25, 2026145Llama 3.1 70B InstructMeta70.6BScore1269Score / $3173Score / GB VRAM~29ScoredSep 25, 2026146Nova Pro 1.0AmazonScore1269Score / $396ScoredSep 25, 2026147Phi 4Microsoft14.7BScore1265Score / $9034Score / GB VRAM140ScoredSep 25, 2026148Mistral Small 3Mistral AI23.6BScore1261Score / $15768Score / GB VRAM~85ScoredSep 25, 2026149Gemma 3n 4BGoogle7.8BScore1259Score / $10496Score / GB VRAM~260ScoredSep 25, 2026150Gemma 3 4BGoogle4.3BScore1254Score / $12540Score / GB VRAM~460ScoredSep 25, 2026151Gemma 2 27BGoogle27.2BScore1248Score / $1919Score / GB VRAM~73ScoredSep 25, 2026152Jamba Large 1.7AI21399BScore1245Score / $156Score / GB VRAM~5ScoredSep 13, 2026153Mistral LargeMistral AIScore1244Score / $207ScoredSep 25, 2026154Nova Lite 1.0AmazonScore1244Score / $5183ScoredSep 25, 2026155Claude 3 HaikuAnthropicScore1232Score / $985ScoredSep 25, 2026156Mixtral 8x22B InstructMistral AI141BScore1228Score / $205Score / GB VRAM~14ScoredSep 25, 2026157Nova Micro 1.0AmazonScore1223Score / $8738ScoredSep 25, 2026158Ministral 8BMistral AIScore1214Score / $11032ScoredSep 25, 2026159Command R (08-2024)CohereScore1207Score / $2012ScoredSep 25, 2026160GPT-3.5 Turbo (older v0613)OpenAIScore1204Score / $602ScoredSep 25, 2026161Llama 3.1 8B InstructMeta8BScore1189Score / $29716Score / GB VRAM~240ScoredSep 25, 2026162Llama 3.2 3B InstructMeta3.2BScore1165Score / $23300Score / GB VRAM~580ScoredSep 25, 2026163Llama 3.2 1B InstructMeta1.2BScore1123Score / $56167Score / GB VRAM~1500ScoredSep 25, 2026

Showing 101–163 of 163 models on this board.