ObviousBench: leaderboard
ObviousBench evaluates language models on text-based tasks designed to catch simple, user-visible mistakes such as arithmetic, spelling, formatting, ordering, and constraint-following errors. The pinned metric is Answer pass³ (%), which measures whether all three sampled attempts are answer-correct. The cited results are from a public v0.2 release with a held-out private set not released. Source: BenchmarkList’s attributed result table.
Metric: Answer pass³ (%). Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 643 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 27B | 100 | #192 |
| 2 | GPT-6 Astra (Max) | 100 | #6 (GPT-6 Astra) |
| 3 | Gemini 3.1 Pro (Preview) (High) | 100 | #64 (Gemini 3.1 Pro (Preview)) |
| 4 | GPT-5.5 (xHigh) | 100 | #33 (GPT-5.5) |
| 5 | Claude Opus 5.5 (Max) | 100 | #3 (Claude Opus 5.5) |
| 6 | GPT-6.1 Sol (Max) | 100 | #7 (GPT-6.1 Sol) |
| 7 | GPT-5 (High) | 100 | #105 (GPT-5) |
| 8 | Gemini 3.5 Flash (High) | 100 | #66 (Gemini 3.5 Flash) |
| 9 | GPT-5 (Medium) | 100 | #105 (GPT-5) |
| 10 | GPT-6 Astra (xHigh) | 100 | #6 (GPT-6 Astra) |
| 11 | Grok 4.5 (High) | 100 | #54 (Grok 4.5) |
| 12 | GPT-6 Astra (High) | 100 | #6 (GPT-6 Astra) |
| 13 | Grok 4.6 (High) | 100 | #29 (Grok 4.6) |
| 14 | Claude Opus 5 (High) | 100 | #16 (Claude Opus 5) |
| 15 | GPT-6 Astra (Medium) | 100 | #6 (GPT-6 Astra) |
No result here: #9 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=obviousbench · How It Works · Data refreshed daily, snapshot 2026-11-05.