SWE-Milestone: leaderboard
SWE-Milestone evaluates coding agents on software-evolution itineraries mined from real repositories as milestone DAGs, requiring sustained multi-milestone implementation; scored by milestone completion with the best harness retained per model.
Metric: Milestone Score (%). Source: swe-milestone.com. Saturation forecast: Around October 2027. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.3 | 58.75 | #39 |
| 2 | GPT-5.6 Sol (Max) | 55.58 | #19 (GPT-5.6 Sol) |
| 3 | Kimi K3 | 51.8 | #30 |
| 4 | GLM-5.2 | 50.31 | #73 |
| 5 | Claude Opus 4.7 (xHigh) | 49.56 | #49 (Claude Opus 4.7) |
| 6 | GPT-5.5 (xHigh) | 47.96 | #31 (GPT-5.5) |
| 7 | GPT-5.4 (xHigh) | 43.26 | #79 (GPT-5.4) |
| 8 | Kimi K2.6 | 42.9 | #104 |
| 9 | Claude Opus 4.6 | 41.99 | #63 |
| 10 | Gemini 3.5 Flash | 39.79 | #60 |
| 11 | GPT-5.3 Codex (xHigh) | 37.75 | #69 (GPT-5.3 Codex) |
| 12 | Claude Sonnet 4.6 | 34.95 | #90 |
| 13 | GPT-5.2 (xHigh) | 33.76 | #110 (GPT-5.2) |
| 14 | GLM-5.1 | 32.61 | #97 |
| 15 | Claude Opus 4.5 | 29.93 | #83 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=swe-milestone · How It Works · Data refreshed daily, snapshot 2026-10-20.