PerfCodeBench - Correct-and-Runnable 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 correct-and-runnable rate: the percentage of tasks whose generated code compiles, runs and passes the task oracle; higher is better.
Metric: Correct-and-runnable rate (%): share of tasks whose generated code compiles, runs and passes the oracle (compilation errors, runtime errors and wrong output all count as failures), 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-5.4 | 71.25 | #135 |
| 2 | Claude Opus 4.5 | 70.55 | #160 |
| 3 | Claude Sonnet 4.5 | 69.36 | #261 |
| 4 | GPT-5 | 61.81 | #178 |
| 5 | DeepSeek V4 Pro | 54.48 | #196 |
| 6 | Gemini 3.1 Flash | 51.83 | #312 |
| 7 | Seed 2.0 Lite | 47.57 | #300 |
| 8 | Qwen 3.6 Max | 47.25 | #193 |
| 9 | DeepSeek V4 Flash | 47.09 | #267 |
| 10 | Qwen 3.6 Plus | 45.2 | #238 |
| 11 | Gemma 4 26B A4B | 45.15 | #590 |
| 12 | Gemma 4 31B | 42.45 | #382 |
| 13 | Kimi K2 | 36.03 | #443 |
| 14 | Qwen 3.6 35B A3B | 32.09 | #481 |
| 15 | Llama 4 Maverick | 29.4 | #1110 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=perfcodebench-correct-and-runnable-rate · How It Works · Data refreshed daily, snapshot 2026-11-21.