AA Omniscience - Software Engineering (SWE) - TypeScript: leaderboard
Metric: Accuracy (%). Source: artificialanalysis.ai. 294 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kimi K3 (Max) | 77.78 |
| 2 | Kimi K3 (Low) | 75.28 |
| 3 | Claude Opus 4.5 (Thinking) | 71.11 |
| 4 | DeepSeek V4 Pro (0813) (Reasoning, Max Effort) | 71.11 |
| 5 | Kimi K2.7 Code | 66.67 |
| 6 | Gemini 3 Pro (Preview) (High) | 64.44 |
| 7 | Claude Opus 4.5 (Non-reasoning) | 64.44 |
| 8 | Kimi K2.5 (Thinking) | 60 |
| 9 | Inkling (xHigh) | 60 |
| 10 | DeepSeek V4 Pro (Reasoning, Max Effort) | 57.78 |
| 11 | DeepSeek V4 Pro (Reasoning, High Effort) | 55.56 |
| 12 | Kimi K2.6 | 54.44 |
| 13 | DeepSeek V4 Flash (Reasoning, Max Effort) | 54.44 |
| 14 | Hy3 | 53.33 |
| 15 | Claude Sonnet 4.5 (Thinking) | 52.22 |
Interactive version: theaggregate.ai/benchmark?slug=aa-omniscience-software-engineering-swe-typescript · How It Works · Data refreshed daily, snapshot 2026-09-05.