LiveCodeBench scores as reported in the LLM Benchmark Matrix. Tests code generation ability on problems from competitive programming platforms, refreshed regularly to prevent contamination.
This definition has not completed editorial and provenance review. Verify its metric, version, source, and evaluation configuration upstream before using it for a model-selection decision.
Continuously updated problem set reduces contamination. Tests algorithmic coding at multiple difficulty levels.
Matrix snapshot — we also have AA's own LiveCodeBench column (which uses a different scale).
Higher is better. Pass rate percentage (0–100%).
Matrix snapshot — we also have AA's own LiveCodeBench column (which uses a different scale).
| # | Model | Score |
|---|---|---|
| 1 | 91.0 | |
| 2 | 90.8 | |
| 3 | 90.8 | |
| 4 | 89.4 | |
| 5 | 89.3 | |
| 6 | 89.3 | |
| 7 | 89.0 | |
| 8 | 88.9 | |
| 9 | 88.7 | |
| 10 | 88.6 | |
| 11 | 88.4 | |
| 12 | 87.8 | |
| 13 | 87.8 | |
| 14 | 87.7 | |
| 15 | 87.5 | |
| 16 | 87.5 | |
| 17 | 87.5 | |
| 18 | 87.1 | |
| 19 | 87.1 | |
| 20 | 86.6 | |
| 21 | 86.5 | |
| 22 | 86.3 | |
| 23 | 86.0 | |
| 24 | 86.0 | |
| 25 | 85.7 | |
| 26 | 85.6 | |
| 27 | 85.6 | |
| 28 | 85.3 | |
| 29 | 85.0 | |
| 30 | 85.0 | |
| 31 | 85.0 | |
| 32 | 85.0 | |
| 33 | 84.9 | |
| 34 | 84.9 | |
| 35 | 84.9 | |
| 36 | 84.5 | |
| 37 | 84.5 | |
| 38 | 84.5 | |
| 39 | 84.5 | |
| 40 | 84.4 | |
| 41 | 84.4 | |
| 42 | 84.4 | |
| 43 | 84.3 | |
| 44 | 83.7 | |
| 45 | 83.7 | |
| 46 | 83.7 | |
| 47 | 83.6 | |
| 48 | 83.6 | |
| 49 | 83.3 | |
| 50 | 83.3 |