AA-Omniscience Index - Software Engineering (SWE) - R: leaderboard
Artificial Analysis's AA-Omniscience Index on the Software Engineering questions about R: the share of correct answers minus the share of incorrect ones, bounded -100 to 100, so abstaining beats guessing wrong. Higher is better; small question sets per language make scores coarse and often negative. Not the accuracy on the frozen AA Omniscience SWE language boards.
Metric: Omniscience Index. Source: artificialanalysis.ai. Saturation forecast: Around 2030. 117 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Carbon (High) | 40 | #1 (Gemini 4 Carbon) |
| 2 | Gemini 4 Argon (High) | 20 | #9 (Gemini 4 Argon) |
| 3 | Muse Wonder Full (Max) | 20 | #12 (Muse Wonder Full) |
| 4 | GPT-6 Astra (xHigh) | 10 | #6 (GPT-6 Astra) |
| 5 | GPT-6 Astra (Max) | 3.33 | #6 (GPT-6 Astra) |
| 6 | GPT-6.1 Sol (Max) | 3.33 | #7 (GPT-6.1 Sol) |
| 7 | Grok 4.7 (Low) | 3.33 | #26 (Grok 4.7) |
| 8 | Kimi K3.1 Pro | 0 | #17 |
| 9 | GPT-6 Astra (High) | 0 | #6 (GPT-6 Astra) |
| 10 | Step 5 Preview | 0 | #53 |
| 11 | GPT-6.1 Terra (Low) | -3.33 | #15 (GPT-6.1 Terra) |
| 12 | GPT-6 Astra (Medium) | -6.67 | #6 (GPT-6 Astra) |
| 13 | GPT-6.1 Sol (Low) | -6.67 | #7 (GPT-6.1 Sol) |
| 14 | GPT-6.1 Terra (Max) | -6.67 | #15 (GPT-6.1 Terra) |
| 15 | Claude Haiku 5.5 (Low) | -6.67 | #43 (Claude Haiku 5.5) |
No result here: #3 Claude Opus 5.5, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=aa-omniscience-index-software-engineering-swe-r · How It Works · Data refreshed daily, snapshot 2026-11-05.