Leaderboards
Benchmarks
Every score carries its origin — leaderboard is independent,vendor is self-reported. History is kept, never overwritten.
Arena Hard Prompts143GPQA Diamond16LiveBench35LiveBench Data Analysis35LiveBench Reasoning35MMLU0MMLU-Pro16
LiveBench Data Analysis
% average over LiveBench table and event-ordering taskshigher is betterProvenanceindependent leaderboard35 modelsscores Jun 25, 2026
1GPT-5.5OpenAI81.58via xHigh8—Jun 25, 2026leaderboard2Claude Fable 5Anthropic80.54via Max effort2—Jun 25, 2026leaderboard3GPT-5.6 SolOpenAI79.84via Max5—Jun 25, 2026leaderboard4GPT-5.4OpenAI79.31via xHigh11—Jun 25, 2026leaderboard5GPT-5.6 TerraOpenAI79.31via Max26—Jun 25, 2026leaderboard6Kimi K3Moonshot AI2.8T78.736~0.045Jun 25, 2026leaderboard7Claude Opus 4.7Anthropic78.26via xHigh effort3—Jun 25, 2026leaderboard8GPT-5.2-CodexOpenAI78.26—Jun 25, 2026leaderboard9GPT-5.2 ChatOpenAI78.16via High11—Jun 25, 2026leaderboard10GPT-5.6 LunaOpenAI78.03via Max59—Jun 25, 2026leaderboard11Claude Sonnet 4.6Anthropic77.95via Thinking Auto Medium effort5—Jun 25, 2026leaderboard12MiniMax M3MiniMax427B76.1779~0.28Jun 25, 2026leaderboard13Claude Opus 5Anthropic74.55via Max effort3—Jun 25, 2026leaderboard14DeepSeek V4 ProDeepSeek1.6T74.5486~0.074Jun 25, 2026leaderboard15Claude Opus 4.5Anthropic74.44via Thinking 64k High effort3—Jun 25, 2026leaderboard16GLM 5.2Z.AI753B73.7449~0.16Jun 25, 2026leaderboard17Grok 4.5xAI73.0412—Jun 25, 2026leaderboard18InklingThinking Machines952B72.78via xHigh18~0.12Jun 25, 2026leaderboard19Muse Spark 1.1Meta72.55via xHigh17—Jun 25, 2026leaderboard20Qwen3.7 MaxQwen71.7919—Jun 25, 2026leaderboard21Claude Sonnet 5Anthropic71.74via xHigh effort7—Jun 25, 2026leaderboard22Grok Build 0.1xAI70.7935—Jun 25, 2026leaderboard23GPT-5.4 MiniOpenAI70.79via xHigh31—Jun 25, 2026leaderboard24Qwen3.6 27BQwen27.8B70.4337~4Jun 25, 2026leaderboard25Qwen3.6 PlusQwen69.9136—Jun 25, 2026leaderboard26Claude Opus 4.6Anthropic69.89via Thinking Auto High effort3—Jun 25, 2026leaderboard27DeepSeek V4 FlashDeepSeek291B68.02386~0.37Jun 25, 2026leaderboard28GPT-5.4 NanoOpenAI67.64via xHigh108—Jun 25, 2026leaderboard29Claude Opus 4.8Anthropic66.03via Max effort3—Jun 25, 2026leaderboard30Kimi K2.6Moonshot AI1.1T65.13via Thinking26~0.098Jun 25, 2026leaderboard31Gemini 3.5 FlashGoogle64.86via High14—Jun 25, 2026leaderboard32Gemini 3.6 FlashGoogle63via High17—Jun 25, 2026leaderboard33Kimi K2.7 CodeMoonshot AI1.1T62.6618~0.094Jun 25, 2026leaderboard34Grok 4.3xAI55.7722—Jun 25, 2026leaderboard35Gemini 3.5 Flash LiteGoogle53.25via High43—Jun 25, 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.