PerfCodeBench - Faster-than-Baseline Rate: leaderboard
Metric: Faster-than-baseline rate (%): share of tasks where the model produces correct code that runs faster than the baseline implementation, 1,854 system-level performance-optimization tasks across C, C++, Go, Java, Python and CUDA; the model returns one replacement source file from the task metadata, interface contract and baseline in one shot; a candidate earns performance credit only if it compiles, runs and passes the task oracle; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 66.72 |
| 2 | Claude Opus 4.5 | 65.75 |
| 3 | Claude Sonnet 4.5 | 62.08 |
| 4 | GPT-5 | 59.55 |
| 5 | DeepSeek V4 Pro | 51.13 |
| 6 | DeepSeek V4 Flash | 44.82 |
| 7 | Seed 2.0 Lite | 43.8 |
| 8 | Qwen 3.6 Max | 42.5 |
| 9 | Qwen 3.6 Plus | 39.75 |
| 10 | Gemma 4 26B A4B | 39.64 |
| 11 | Gemma 4 31B | 37.65 |
| 12 | Kimi K2 | 32.9 |
| 13 | Qwen 3.6 35B A3B | 27.94 |
| 14 | Llama 4 Maverick | 24.38 |
| 15 | Gemini 3.1 Pro (Preview) | 23.73 |
Interactive version: theaggregate.ai/benchmark?slug=perfcodebench-faster-than-baseline-rate · How It Works · Data refreshed daily, snapshot 2026-10-07.