SWE-2 Coding Benchmark
How SWE-2 ranks on the LLM Coding Leaderboard, shown next to the three models directly above and below it, with the top three models included for reference.
- Last evaluated: 2 weeks ago (October 1, 2026), with Devin.
| # | Model | Total points (max 80) |
Avg cost per prompt |
Avg time per prompt |
Video | Tested with | Points per project | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
Laravel Code Quality
(max 20) |
React-TS Code Quality
(max 20) |
Bug Finding Test
(max 20) |
CSV Import (PHP) | Offline Sync (PHP) | Bank Feed (Dart / Flutter) | Shipping Quotes (Go) | |||||||
| 1 | GPT-6.1-Sol (High) | 68.1 | $0.22 | 07:26 | Codex CLI | 19.1 | 19 | 10 | 5 | 5 | 5 | 5 | |
| 2 | Opus 5.5 (High) | 67.41 | $0.79 | 03:10 | Claude Code | 18.5 | 19.33 | 9.58 | 5 | 5 | 5 | 5 | |
| 3 | GPT-6.1-Sol (Medium) | 67.09 | $0.14 | 03:53 | Codex CLI | 18.63 | 18.67 | 9.79 | 5 | 5 | 5 | 5 | |
| ⋯ 12 places not shown ⋯ | |||||||||||||
| 16 | Haiku 5.5 (Medium) | 58.2 | $0.03 | 02:37 | Claude Code | 17.32 | 17 | 6.88 | 4 | 4.5 | 3.5 | 5 | |
| 17 | Space Bunny (Max) | 57.19 | N/A | 19:19 | OpenCode | 17.61 | 16.5 | 4.58 | 5 | 4.5 | 4 | 5 | |
| 18 | Gemini 4 Argon (High) | 57.09 | N/A | 09:41 | Devin | 17.15 | 18.5 | 7.19 | 2.5 | 3.75 | 5 | 3 | |
| 19 | SWE-2 (Max) | 57.07 | N/A | 11:59 | Devin | 17.5 | 17.83 | 7.29 | 4 | 2.45 | 4 | 4 | |
| 20 | Kimi K3 (High) | 56.89 | $0.63 | 12:49 | OpenCode | 18.4 | 17.5 | 6.04 | 4.5 | 3.25 | 3.2 | 4 | |
| 21 | Mistral Large 4 (High) | 56.86 | $0.34 | 14:33 | OpenCode | 18.13 | 19.33 | 3.65 | 3.5 | 3.75 | 3.5 | 5 | |
| 22 | Grok 4.6 (High) | 56.71 | $0.77 | 09:19 | Cursor | 17.4 | 18 | 6.56 | 4 | 3.75 | 4.5 | 2.5 | |
Tutorials about SWE-2
Video
· Sep 15, 2026
I Tested NEW SWE-2 Model in Devin on 24 Coding Prompts
How to read this
- Each score measures a model-and-harness configuration, not the model in isolation.
- Prices are calculated with API costs. Models tested on a subscription plan show N/A.
- Full methodology and scoring formulas are explained in this article.
See every tested model side by side on the full LLM Coding Leaderboard.
