Leaderboards
Benchmarks
Every score carries its origin — leaderboard is independent,vendor is self-reported. History is kept, never overwritten.
LiveCodeBench
% pass on contamination-free competitive programming problemshigher is betterProvenanceindependent leaderboard14 modelsscores Aug 3, 2026
1o4 Mini HighOpenAI87.3via High20—Aug 3, 2026leaderboard2o3OpenAI84.7via High11—Aug 3, 2026leaderboard3o4 MiniOpenAI84.5via Medium19—Aug 3, 2026leaderboard4R1 0528DeepSeek685B84.439~0.2Aug 3, 2026leaderboard5Gemini 2.5 ProGoogle84.317—Aug 3, 2026leaderboard6Qwen3 235B A22B Instruct 2507Qwen235B80.4146~0.54Aug 3, 2026leaderboard7o3 MiniOpenAI77.7via High18—Aug 3, 2026leaderboard8Gemini 2.5 FlashGoogle76.261—Aug 3, 2026leaderboard9Claude Opus 4Anthropic62.41—Aug 3, 2026leaderboard10DeepSeek V3DeepSeek685B49.6560Aug 3, 2026leaderboard11GPT-4o (2024-08-06)OpenAI38.34—Aug 3, 2026leaderboard12GPT-4 TurboOpenAI37.31—Aug 3, 2026leaderboard13GPT-4o-mini (2024-07-18)OpenAI35.559—Aug 3, 2026leaderboard14Claude 3 HaikuAnthropic22.518—Aug 3, 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.