SWE-Bench Pro V2 Full
Repository issue resolution on the full public V2 set. Models must edit an existing codebase to satisfy the issue’s verification tests.
How it is run
Published September 2026 V2 table. Resolved-task percentage; each row retains its agent scaffold and effort setting.
What it cannot tell you
Full and HARD are distinct task sets. Claude Code, Codex and mini-swe-agent results use different scaffolds and are not a controlled model-only comparison.
- Leader
- Opus 5 (Claude Code) xhigh99.40% · Scale AI / SEAL results, grouped here by release
- Top-ten spread
- 9.52%Leader minus tenth place, or last place in smaller fields.
- Median
- 95.55%of the 10 rows published
- No published result
- 36of 46 models on this site
Scale AI / SEAL results by release
V2 FULL field, as published. Different agent scaffolds and effort settings are retained per row; this is a model-and-agent comparison.
| # | Model | Score | Effort |
|---|---|---|---|
| 1 | Opus 5 (Claude Code) xhigh | 99.40% | (Claude Code) xhigh |
| 2 | Fable 5.1 (Claude Code) high | 99.10% | (Claude Code) high |
| 3 | Kimi-K3 (mini-swe-agent) max | 97.70% | (mini-swe-agent) max |
| 4 | GPT-6-Astra (Codex) high | 96.90% | (Codex) high |
| 5 | GLM-5.3 (mini-swe-agent) max | 95.60% | (mini-swe-agent) max |
| 6 | GPT-5.6-sol (Codex) xhigh | 95.50% | (Codex) xhigh |
| 7 | Gemini 3.8 Flash (mini-swe-agent) high | 94.86% | (mini-swe-agent) high |
| 8 | Sonnet 5 (Claude Code) xhigh | 93.15% | (Claude Code) xhigh |
| 9 | GPT-5.6-Terra (Codex) xhigh | 92.37% | (Codex) xhigh |
| 10 | Inkling (mini-swe-agent) xhigh | 89.88% | (mini-swe-agent) xhigh |
No published result tracked
36 of the 46 models on this site have no published number on this benchmark, so their cells in the home table stay blank rather than reading as zero.
Read this with
Other versions of the same evaluation. Versions are not comparable: the task set changes between them.
Sources
Every page used for this benchmark, read on 17 October 2026.
- Benchmark site or the page that publishes itlabs.scale.com/leaderboard/swe_bench_pro_public_v2
- Independent resultslabs.scale.com/leaderboard/swe_bench_pro_public_v2