PDB-Single - Unit Test Pass Rate: leaderboard
The Precise Debugging Benchmark (PDB, 2026) turns coding datasets into debugging tasks with verified injected bugs and checks whether a model fixes them with minimal targeted edits instead of regenerating the program. This board is the share of debugged programs that pass their unit tests on the PDB-Single-Full set, in percent; higher is better.
Metric: Unit-test pass rate (%) of the debugged programs on the PDB-Single-Full set; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 8 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | DeepSeek V3.2 | Thinking | 84.7 | #419 |
| 2 | Gemini 2.5 Pro | 83.8 | #276 | |
| 3 | GPT-5.1 Codex | 82 | #295 | |
| 4 | Claude Sonnet 4.5 | Thinking | 81.9 | #261 |
| 5 | Kimi K2 (Thinking) | 80.8 | ||
| 6 | Qwen 3 Coder 480B A35B | 77.4 | #709 | |
| 7 | Kimi K2 | 73 | #443 | |
| 8 | Grok Code Fast 1 | 67.1 | #703 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=pdb-single-unit-test-pass-rate · How It Works · Data refreshed daily, snapshot 2026-11-21.