LiveBench Spatial: leaderboard
Reasoning task in LiveBench built from hand written questions that ask the model to deduce how common two and three dimensional shapes intersect and how they are oriented. Answers are checked against an objective ground truth with no LLM judge, and LiveBench rotates its questions over time. The score runs from 0 to 100 and higher is better.
Metric: Score. Source: livebench.ai. Saturation forecast: Estimated already saturated. 71 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.6 27B | 100 | #170 |
| 2 | Kimi K3 | 100 | #30 |
| 3 | Qwen 3.8 Max | 100 | #35 |
| 4 | Grok 4.5 | 100 | #48 |
| 5 | GPT-5.6 Sol (Max) | 100 | #19 (GPT-5.6 Sol) |
| 6 | Claude Opus 5 (Max) | 100 | #12 (Claude Opus 5) |
| 7 | Gemini 3.8 Flash (High) | 100 | #21 (Gemini 3.8 Flash) |
| 8 | GPT-5.6 Terra (Max) | 100 | #41 (GPT-5.6 Terra) |
| 9 | Qwen3.8-Flash-Next | 100 | #72 |
| 10 | Gemini 3.7 Flash (High) | 100 | #23 (Gemini 3.7 Flash) |
| 11 | Gemini 3.6 Flash (High) | 100 | #51 (Gemini 3.6 Flash) |
| 12 | Claude Fable 5.1 (Max) | 100 | #9 (Claude Fable 5.1) |
| 13 | Claude Opus 4.7 (xHigh) | 100 | #49 (Claude Opus 4.7) |
| 14 | Muse Spark 1.3 (xHigh) | 100 | #24 (Muse Spark 1.3) |
| 15 | Claude Sonnet 5 (xHigh) | 100 | #57 (Claude Sonnet 5) |
No result here: #8 Gemini 4 Argon, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=livebench-spatial · How It Works · Data refreshed daily, snapshot 2026-10-20.