Competitive coding
LiveCodeBench
Coding problems published so recently the AI can't have seen them in training — a contamination-free test of raw programming skill. Higher is better.
Rankings
Higher is betterLiveCodeBench — frequently asked questions
- What is LiveCodeBench?
- Coding problems published so recently the AI can't have seen them in training — a contamination-free test of raw programming skill. Higher is better.
- Which AI model scores highest on LiveCodeBench?
- DeepSeek-V4-Pro by DeepSeek holds the best LiveCodeBench result among tracked models, at 93.5% (released Apr 24 2026). Higher scores are better on this benchmark.
- What are the top 4 models on LiveCodeBench?
- 1. DeepSeek-V4-Pro (DeepSeek) — 93.5%; 2. DeepSeek-V4-Flash (DeepSeek) — 91.6%; 3. Kimi K2.6 (Moonshot AI) — 89.6%; 4. Kimi K2.5 (Moonshot AI) — 85%.
- How many models have a published LiveCodeBench score?
- 4 tracked models have a published LiveCodeBench score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.