GameDevBench: leaderboard
GameDevBench evaluates LLM agents on game-development tasks in the Godot engine using text plus multimodal assets such as shaders, sprites, animations, and visual game scenes. The tasks cover 2D graphics, 3D graphics, UI, and gameplay logic. The pinned metric is Pass@1, measuring the share of tasks solved on the first attempt. Source: BenchmarkList’s attributed result table.
Metric: Pass@1 (%). Source: benchmarklist.com. Saturation forecast: Around May 2027. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (xHigh) | 70.6 | #3 (Claude Opus 5.5) |
| 2 | GPT-6.1 Sol (xHigh) | 70 | #7 (GPT-6.1 Sol) |
| 3 | GPT-6 Astra (High) | 68.8 | #6 (GPT-6 Astra) |
| 4 | Claude Opus 5.5 (High) | 68.2 | #3 (Claude Opus 5.5) |
| 5 | Claude Fable 5 (xHigh) | 67.3 | #20 (Claude Fable 5) |
| 6 | GPT-5.6 Sol (xHigh) | 63.7 | #22 (GPT-5.6 Sol) |
| 7 | GPT-5.6 Sol (High) | 63.1 | #22 (GPT-5.6 Sol) |
| 8 | GPT-5.6 Sol (Medium) | 58.6 | #22 (GPT-5.6 Sol) |
| 9 | Kimi K3 | 58 | #35 |
| 10 | Claude Opus 4.8 | 55.9 | #39 |
| 11 | GPT-5.5 | 54.7 | #33 |
| 12 | Gemini 3 Pro (Preview) | 53.8 | #71 |
| 13 | GPT-5.4 | 52 | #83 |
| 14 | Gemini 3 Flash (Preview) | 46.8 | #90 |
| 15 | GPT-5.4 Mini | 43.2 | #228 |
No result here: #9 Gemini 4 Argon, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=gamedevbench · How It Works · Data refreshed daily, snapshot 2026-11-05.