Leaderboards
Benchmarks
Every score carries its origin — leaderboard is independent,vendor is self-reported. History is kept, never overwritten.
Arena Coding
Arena Score on coding prompts within the text arena — not a percentagehigher is betterProvenanceindependent leaderboard100 modelsscores Aug 2, 2026
1Claude Fable 5Anthropic1552.827831—Aug 2, 2026leaderboard2Claude Opus 4.6Anthropic1547.865862—Aug 2, 2026leaderboard3Claude Opus 4.7Anthropic1547.031462—Aug 2, 2026leaderboard4Claude Opus 5Anthropic1532.0375via High61—Aug 2, 2026leaderboard5Kimi K3Moonshot AI2.8T1531.4745via Max109~0.87Aug 2, 2026leaderboard6Muse Spark 1.1Meta1529.1558360—Aug 2, 2026leaderboard7Claude Opus 4.8Anthropic1527.693161—Aug 2, 2026leaderboard8GPT-5.6 SolOpenAI1527.6262via xHigh102—Aug 2, 2026leaderboard9Claude Sonnet 4.6Anthropic1526.9084102—Aug 2, 2026leaderboard10Qwen3.7 MaxQwen1525.3305407—Aug 2, 2026leaderboard11Claude Opus 4.5Anthropic1522.973161—Aug 2, 2026leaderboard12Gemini 3.6 FlashGoogle1521.4722406—Aug 2, 2026leaderboard13Gemini 3.1 Pro Preview Custom ToolsGoogle1520.7673253—Aug 2, 2026leaderboard14GPT-5.6 TerraOpenAI1520.238via xHigh507—Aug 2, 2026leaderboard15Grok 4.5xAI1519.6655253—Aug 2, 2026leaderboard16GLM 5.1Z.AI754B1518.4089507~3.2Aug 2, 2026leaderboard17MiMo-V2.5-ProXiaomi1T1518.33862182~2.4Aug 2, 2026leaderboard18GPT-5.2 ChatOpenAI1515.239216—Aug 2, 2026leaderboard19Kimi K2.6Moonshot AI1.1T1514.8204611~2.3Aug 2, 2026leaderboard20GPT-5.5OpenAI1514.3296151—Aug 2, 2026leaderboard21GPT-5.4OpenAI1513.9884202—Aug 2, 2026leaderboard22Claude Sonnet 4.5Anthropic1513.1141101—Aug 2, 2026leaderboard23Qwen3.7 PlusQwen1509.18751179—Aug 2, 2026leaderboard24Grok 4.20 Multi-AgentxAI1508.9917604—Aug 2, 2026leaderboard25Grok 4.20xAI1508.0767603—Aug 2, 2026leaderboard26Qwen3.6 Max PreviewQwen1507.974245—Aug 2, 2026leaderboard27Gemini 3.5 FlashGoogle1507.5386via Medium335—Aug 2, 2026leaderboard28Gemini 3.5 Flash LiteGoogle1506.18791205—Aug 2, 2026leaderboard29GLM 5.2Z.AI753B1505.7178via Max1004~3.2Aug 2, 2026leaderboard30Claude Opus 4.1Anthropic1505.135520—Aug 2, 2026leaderboard31Kimi K2.5Moonshot AI1.1T1504.7115752~2.3Aug 2, 2026leaderboard32DeepSeek V4 ProDeepSeek1.6T1502.19911727~1.5Aug 2, 2026leaderboard33GPT-5.6 LunaOpenAI1501.3694via xHigh1137—Aug 2, 2026leaderboard34Kimi K2 ThinkingMoonshot AI1.1T1501.1202via Thinking600~2.3Aug 2, 2026leaderboard35Hy3Tencent299B1500.84052843~8Aug 2, 2026leaderboard36Gemma 4 31BGoogle32.7B1498.12744406~73Aug 2, 2026leaderboard37MiniMax M3MiniMax427B1497.75941560~5.6Aug 2, 2026leaderboard38GLM 5Z.AI754B1497.4049780~3.2Aug 2, 2026leaderboard39InklingThinking Machines952B1497.1357370~2.5Aug 2, 2026leaderboard40GPT-5.4 MiniOpenAI1497.045via High665—Aug 2, 2026leaderboard41GPT-5.3 ChatOpenAI1496.7269107—Aug 2, 2026leaderboard42Qwen3.6 PlusQwen1494.673766—Aug 2, 2026leaderboard43MiMo-V2.5Xiaomi311B1491.26657~7.6Aug 2, 2026leaderboard44Qwen3.5 397B A17BQwen403B1490.8235637~5.9Aug 2, 2026leaderboard45GLM 5V TurboZ.AI1488.9544372—Aug 2, 2026leaderboard46Grok 4.3xAI1488.1078595—Aug 2, 2026leaderboard47GLM 4.7Z.AI358B1485.4045849~6.6Aug 2, 2026leaderboard48DeepSeek V4 FlashDeepSeek291B1483.51158410~8.1Aug 2, 2026leaderboard49Qwen3 MaxQwen1481.7442380—Aug 2, 2026leaderboard50Gemma 4 26B A4BGoogle26.5B1480.74424936~89Aug 2, 2026leaderboard51MiniMax M2.7MiniMax229B1479.44861541~10Aug 2, 2026leaderboard52Claude Haiku 4.5Anthropic1479.011296—Aug 2, 2026leaderboard53Mistral Medium 3.5Mistral AI128B1478.8609197~18Aug 2, 2026leaderboard54GPT-5.1OpenAI1474.3073295—Aug 2, 2026leaderboard55Qwen3 235B A22B Instruct 2507Qwen235B1472.54972677~9.9Aug 2, 2026leaderboard56DeepSeek V3.2DeepSeek685B1469.59294728~3.4Aug 2, 2026leaderboard57Mistral Large 3 2512Mistral AI1468.0022979—Aug 2, 2026leaderboard58Gemini 2.5 ProGoogle1465.2339293—Aug 2, 2026leaderboard59Qwen3 VL 235B A22B InstructQwen236B1464.96241665~9.9Aug 2, 2026leaderboard60DeepSeek V3.2 ExpDeepSeek685B1464.93953573~3.4Aug 2, 2026leaderboard61Claude Opus 4Anthropic1464.487920—Aug 2, 2026leaderboard62R1 0528DeepSeek685B1464.4555681~3.4Aug 2, 2026leaderboard63GPT-5.4 NanoOpenAI1460.7738via High2337—Aug 2, 2026leaderboard64Hy3 previewTencent299B1459.47336950~7.7Aug 2, 2026leaderboard65o3OpenAI1459.3502182—Aug 2, 2026leaderboard66Qwen3.5-122B-A10BQwen125B1458.9754701~18Aug 2, 2026leaderboard67GLM 4.6Z.AI357B1458.951834~6.5Aug 2, 2026leaderboard68Gemini 3.1 Flash Lite PreviewGoogle1457.39941943—Aug 2, 2026leaderboard69Qwen3 Coder 480B A35BQwen480B1456.83131457~4.8Aug 2, 2026leaderboard70GPT-4.1OpenAI1456.0136182—Aug 2, 2026leaderboard71Qwen3 VL 235B A22B ThinkingQwen236B1454.6093via Thinking559~9.8Aug 2, 2026leaderboard72GLM 4.5Z.AI358B1454.3498661~6.4Aug 2, 2026leaderboard73Step 3.5 FlashStepFun199B1450.38084835~12Aug 2, 2026leaderboard74Qwen3.5-27BQwen27.8B1449.642929~83Aug 2, 2026leaderboard75Qwen3 Next 80B A3B InstructQwen81.3B1445.49281853~28Aug 2, 2026leaderboard76MiniMax M2.5MiniMax229B1444.23711605~10Aug 2, 2026leaderboard77Qwen3 235B A22B Thinking 2507Qwen235B1441.4854964~9.7Aug 2, 2026leaderboard78Qwen3 30B A3B Instruct 2507Qwen30.5B1439.3357456~75Aug 2, 2026leaderboard79DeepSeek V3.1 TerminusDeepSeek685B1439.00091515~3.3Aug 2, 2026leaderboard80Qwen3.5-FlashQwen1437.24625528—Aug 2, 2026leaderboard81Claude Sonnet 5Anthropic1435.1588144—Aug 2, 2026leaderboard82Qwen3.5-35B-A3BQwen36B1434.87551435~63Aug 2, 2026leaderboard83o3 Mini HighOpenAI1434.6865via High326—Aug 2, 2026leaderboard84GPT-4.1 MiniOpenAI1433.5678896—Aug 2, 2026leaderboard85o1OpenAI1433.00424—Aug 2, 2026leaderboard86o4 MiniOpenAI1432.4818326—Aug 2, 2026leaderboard87GPT-5 MiniOpenAI1431.0336via High1431—Aug 2, 2026leaderboard88DeepSeek V3 0324DeepSeek685B1428.74441587~3.3Aug 2, 2026leaderboard89GLM 4.5 AirZ.AI111B1426.2061678~20Aug 2, 2026leaderboard90Gemini 2.5 FlashGoogle1423.59651139—Aug 2, 2026leaderboard91GLM 4.7 FlashZ.AI31.2B1423.4292355975Aug 2, 2026leaderboard92Qwen3 Next 80B A3B ThinkingQwen81.3B1420.7329via Thinking1821~28Aug 2, 2026leaderboard93GLM 4.6VZ.AI108B1418.60911576~21Aug 2, 2026leaderboard94MiniMax M1MiniMax1416.0966644—Aug 2, 2026leaderboard95o3 MiniOpenAI1415.9383322—Aug 2, 2026leaderboard96Trinity Large ThinkingArcee AI399B1414.3271via Thinking1768~5.6Aug 2, 2026leaderboard97Ling-2.6-flashinclusionAI107B1412.146647072~21Aug 2, 2026leaderboard98Nemotron 3 SuperNVIDIA124B1409.7023524~18Aug 2, 2026leaderboard99Qwen3 32BQwen32.8B1407.24695026~68Aug 2, 2026leaderboard100GLM 4.5VZ.AI108B1404.2302780~21Aug 2, 2026leaderboard
A tilde in Score / GB VRAM means the memory figure behind it is calculated from the parameter count rather than read from a published file. Those ratios are shown to two significant figures, and rows within a few percent of each other are not distinguishable.