PerfCodeBench - Reference-Beating Rate: leaderboard
PerfCodeBench (2026) tests whether LLMs can optimize real system-level code: 1,854 executable tasks over C, C++, Go, Java, Python and CUDA, where the model rewrites a baseline and is timed against a baseline and an expert reference. This board is the percentage of comparable tasks where the correct code matches or beats the expert reference implementation; higher is better.
Metric: Reference-beating rate (%): share of comparable tasks where the correct code matches or beats the reference 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 February 2027. 20 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 61.6 | #178 |
| 2 | Kimi K2.6 | 52.13 | #192 |
| 3 | GPT-5.4 | 51.82 | #135 |
| 4 | DeepSeek V4 Pro | 47.95 | #196 |
| 5 | Claude Opus 4.5 | 43.23 | #160 |
| 6 | DeepSeek V4 Flash | 43.05 | #267 |
| 7 | Seed 2.0 Lite | 40.68 | #300 |
| 8 | Gemini 3.1 Pro (Preview) | 40.09 | #99 |
| 9 | Qwen 3.6 27B | 39.6 | #351 |
| 10 | Qwen 3.6 Max | 37.19 | #193 |
| 11 | Claude Sonnet 4.5 | 32.81 | #261 |
| 12 | Qwen 3.6 Plus | 29.7 | #238 |
| 13 | Kimi K2 | 27.58 | #443 |
| 14 | Qwen 3.6 35B A3B | 22.51 | #481 |
| 15 | Gemini 3.1 Flash | 22.02 | #312 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=perfcodebench-reference-beating-rate · How It Works · Data refreshed daily, snapshot 2026-11-21.