Leaderboards

Benchmarks

Every score carries its origin.

Reasoning374Coding337Math221Agentic375Chat636Speech778
Arena Hard Prompts168GPQA Diamond16LiveBench58LiveBench Data Analysis58LiveBench Reasoning58MMLU-Pro16

Arena Hard Prompts

Arena Score on the hard-prompt subset — not a percentage
Provenance100 of 168 modelsscores Aug 10, 2026 – Sep 25, 2026
1Claude Opus 5.5AnthropicScore1541Score / $77ScoredSep 25, 20262Claude Fable 5AnthropicScore1532Score / $31ScoredSep 13, 20263Claude Opus 4.6AnthropicScore1527Score / $61ScoredSep 25, 20264Claude Fable 5.1AnthropicScore1520Score / $30ScoredSep 25, 20265Claude Opus 4.7AnthropicScore1520Score / $61ScoredSep 25, 20266Kimi K3Moonshot AI2.8TScore1517Score / $155Score / GB VRAM~0.87ScoredSep 25, 20267Claude Opus 5AnthropicScore1516Score / $61ScoredSep 25, 20268Muse Spark 1.3MetaScore1516Score / $357ScoredSep 25, 20269Gemini 3.8 FlashGoogleScore1516Score / $808ScoredSep 25, 202610Muse Spark 1.2MetaScore1511Score / $356ScoredSep 25, 202611Muse Spark 1.1MetaScore1511Score / $356ScoredSep 25, 202612Gemini 3.7 FlashGoogleScore1509Score / $805ScoredSep 25, 202613GPT-5.6 SolOpenAIScore1509Score / $302ScoredSep 25, 202614Gemini 3.1 Pro Preview Custom ToolsGoogleScore1507Score / $251ScoredSep 25, 202615GLM 5.3Z.aiScore1506Score / $1321ScoredSep 25, 202616Claude Opus 4.8AnthropicScore1504Score / $60ScoredSep 25, 202617Claude Sonnet 4.6AnthropicScore1503Score / $100ScoredSep 25, 202618Gemini 3.6 FlashGoogleScore1503Score / $802ScoredAug 10, 202619GPT-6 AstraOpenAIScore1502Score / $60ScoredSep 25, 202620DeepSeek V4.1 FlashDeepSeek763BScore1501Score / $3753Score / GB VRAM~3.1ScoredSep 25, 202621Qwen3.8 MaxQwenScore1501Score / $303ScoredSep 25, 202622GLM 5.3 FlashZ.ai321BScore1499Score / $5998Score / GB VRAM~7.4ScoredSep 25, 202623GPT-5.5OpenAIScore1499Score / $100ScoredSep 25, 202624Claude Opus 4.5AnthropicScore1498Score / $60ScoredSep 25, 202625GPT-5.2 ChatOpenAIScore1497Score / $214ScoredSep 25, 202626Qwen3.7 MaxQwenScore1496Score / $399ScoredSep 25, 202627GLM 5.2Z.ai753BScore1496Score / $1644Score / GB VRAM~3.1ScoredSep 25, 202628MiMo-V2.5-ProXiaomi1TScore1496Score / $2456Score / GB VRAM~2.3ScoredSep 25, 202629Gemini 3.5 FlashGoogleScore1495Score / $332ScoredSep 25, 202630Grok 4.5xAIScore1489Score / $248ScoredSep 25, 202631GLM 5.1Z.ai754BScore1488Score / $491Score / GB VRAM~3.1ScoredSep 25, 202632GPT-5.6 TerraOpenAIScore1488Score / $248ScoredSep 25, 202633Grok 4.20xAIScore1487Score / $595ScoredSep 25, 202634GPT-5.4OpenAIScore1487Score / $198ScoredSep 25, 202635Kimi K2.6Moonshot AI1.1TScore1485Score / $813Score / GB VRAM~2.2ScoredSep 25, 202636GPT-6 SolOpenAIScore1485Score / $297ScoredSep 25, 202637Grok 4.20 Multi-AgentxAIScore1483Score / $593ScoredSep 25, 202638DeepSeek V4 ProDeepSeek1.6TScore1483Score / $947Score / GB VRAM~1.5ScoredSep 25, 202639Claude Sonnet 4.5AnthropicScore1483Score / $99ScoredSep 25, 202640Qwen3.6 Max PreviewQwenScore1482Score / $241ScoredSep 25, 202641MiMo-V2.6-FlashXiaomiScore1481Score / $5568ScoredSep 25, 202642Grok 4.6xAIScore1478Score / $246ScoredSep 25, 202643Claude Opus 4.1AnthropicScore1478Score / $20ScoredSep 25, 202644GLM 5Z.ai754BScore1478Score / $770Score / GB VRAM~3.1ScoredSep 25, 202645Hy3Tencent299BScore1476Score / $2796Score / GB VRAM~7.8ScoredSep 25, 202646Gemma 4 31BGoogle31.3BScore1476Score / $4472Score / GB VRAM~75ScoredSep 25, 202647Gemini 3.5 Flash LiteGoogleScore1475Score / $1180ScoredSep 25, 202648GPT-5.6 LunaOpenAIScore1475Score / $2458ScoredSep 25, 202649Qwen3.7 PlusQwenScore1474Score / $1151ScoredSep 25, 202650GPT-5.3 ChatOpenAIScore1473Score / $105ScoredSep 25, 202651Kimi K2 ThinkingMoonshot AI1.1TScore1472Score / $589Score / GB VRAM~2.2ScoredSep 13, 202652GPT-5.4 MiniOpenAIScore1469Score / $653ScoredSep 25, 202653GPT-6 LunaOpenAIScore1468Score / $5872ScoredSep 25, 202654Qwen3.6 PlusQwenScore1468Score / $753ScoredSep 25, 202655Qwen3.5 397B A17BQwen403BScore1463Score / $625Score / GB VRAM~5.8ScoredSep 25, 202656MiniMax M3MiniMax427BScore1463Score / $1524Score / GB VRAM~5.4ScoredSep 25, 202657InklingThinking Machines952BScore1462Score / $361Score / GB VRAM~2.4ScoredSep 25, 202658Qwen3.8 27BQwen27.8BScore1462Score / $981Score / GB VRAM~83ScoredSep 25, 202659MiMo-V2.5Xiaomi311BScore1462Score / $6142Score / GB VRAM~7.5ScoredSep 25, 202660GLM 4.7Z.ai358BScore1462Score / $835Score / GB VRAM~6.5ScoredSep 25, 202661Gemma 4 26B A4BGoogle25.8BScore1461Score / $6640Score / GB VRAM~90ScoredSep 25, 202662Kimi K2.5Moonshot AI1.1TScore1460Score / $649Score / GB VRAM~2.2ScoredSep 25, 202663DeepSeek V4 FlashDeepSeek291BScore1459Score / $17426Score / GB VRAM~8ScoredSep 25, 202664Qwen3 MaxQwenScore1459Score / $374ScoredSep 25, 202665Gemini 2.5 ProGoogleScore1458Score / $292ScoredSep 25, 202666Grok 4.7xAIScore1457Score / $243ScoredSep 25, 202667Grok 4.3xAIScore1456Score / $582ScoredSep 25, 202668GPT-5.1OpenAIScore1456Score / $291ScoredSep 25, 202669GLM 5V TurboZ.aiScore1452Score / $363ScoredSep 25, 202670GPT-5OpenAIScore1449Score / $145ScoredSep 25, 202671Qwen3 235B A22B Instruct 2507Qwen235BScore1447Score / $4135Score / GB VRAM~9.8ScoredSep 25, 202672DeepSeek V3.2DeepSeek685BScore1447Score / $4673Score / GB VRAM~3.3ScoredSep 25, 202673DeepSeek V3.2 ExpDeepSeek685BScore1446Score / $3528Score / GB VRAM~3.3ScoredSep 25, 202674Muse Glimmer 30BMeta29.8BScore1445Score / $1314Score / GB VRAM~77ScoredSep 25, 202675Gemini 3.1 Flash Lite PreviewGoogleScore1445Score / $1927ScoredSep 25, 202676Mistral Medium 3.5Mistral AI128BScore1445Score / $193Score / GB VRAM~18ScoredSep 25, 202677GLM 4.6Z.ai357BScore1442Score / $824Score / GB VRAM~6.4ScoredSep 25, 202678Claude Haiku 4.5AnthropicScore1441Score / $288ScoredSep 25, 202679o3OpenAIScore1441Score / $180ScoredSep 25, 202680MiniMax M2.7MiniMax229BScore1441Score / $1715Score / GB VRAM~10ScoredSep 25, 202681Qwen3 VL 235B A22B InstructQwen236BScore1438Score / $1634Score / GB VRAM~9.7ScoredSep 25, 202682Claude Opus 4AnthropicScore1436Score / $19ScoredSep 25, 202683Hy3 previewTencent299BScore1434Score / $2390Score / GB VRAM~7.6ScoredSep 25, 202684GLM 4.5Z.ai358BScore1433Score / $652Score / GB VRAM~6.3ScoredSep 25, 202685R1 0528DeepSeek685BScore1433Score / $666Score / GB VRAM~3.3ScoredSep 25, 202686Qwen3.5-122B-A10BQwen125BScore1433Score / $689Score / GB VRAM~18ScoredSep 25, 202687GPT-4.1OpenAIScore1432Score / $179ScoredSep 25, 202688DeepSeek V3.1DeepSeek685BScore1432Score / $1508Score / GB VRAM~3.3ScoredSep 25, 202689Mistral Large 3 2512Mistral AIScore1431Score / $954ScoredSep 25, 202690Inkling SmallThinking Machines266BScore1428Score / $1190Score / GB VRAM~8.5ScoredSep 25, 202691Qwen3.5-27BQwen27.8BScore1427Score / $915Score / GB VRAM~81ScoredSep 25, 202692DeepSeek V3.1 TerminusDeepSeek685BScore1424Score / $1424Score / GB VRAM~3.3ScoredSep 25, 202693GPT-5.4 NanoOpenAIScore1421Score / $2274ScoredSep 25, 202694Qwen3 Next 80B A3B InstructQwen81.3BScore1420Score / $1820Score / GB VRAM~28ScoredSep 25, 202695Gemini 2.5 FlashGoogleScore1419Score / $1135ScoredSep 25, 202696Claude Sonnet 4AnthropicScore1419Score / $95ScoredSep 25, 202697o1OpenAIScore1418Score / $24ScoredSep 25, 202698Qwen3 235B A22B Thinking 2507Qwen235BScore1418Score / $616Score / GB VRAM~9.6ScoredSep 25, 202699MiniMax M2.5MiniMax229BScore1417Score / $1492Score / GB VRAM~9.8ScoredSep 25, 2026100Qwen3 VL 235B A22B ThinkingQwen236BScore1417Score / $359Score / GB VRAM~9.5ScoredSep 25, 2026

Showing 1–100 of 168 models on this board.