Google Gemini 3 Deep Think - Humanity's Last Exam (no tools): leaderboard
The "Humanity's Last Exam (no tools)" subtask of Google Gemini 3 Deep Think. A benchmark result reported for Google Gemini 3 Deep Think, as self-reported by the model developer.
Metric: Score (%). Source: storage.googleapis.com. Source: qwen.ai. Saturation forecast: Around 2029. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Deep Think | 48.4 | #27 |
| 2 | Qwen 3.7 Max | 41.4 | #75 |
| 3 | Claude Opus 4.6 (Thinking) | 40 | #63 (Claude Opus 4.6) |
| 4 | DeepSeek V4 Pro (Reasoning, Max Effort) | 37.7 | #102 (DeepSeek V4 Pro) |
| 5 | Gemini 3 Pro (Preview) | 37.5 | #66 |
| 6 | Kimi K2.6 (Thinking) | 36.4 | #104 (Kimi K2.6) |
| 7 | GLM-5.1 (Thinking) | 34.7 | #97 (GLM-5.1) |
| 8 | GPT-5.2 (xHigh) | 34.5 | #110 (GPT-5.2) |
| 9 | Qwen 3.6 Plus | 28.8 | #120 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=google-gemini-3-deep-think-humanity-s-last-exam-no-tools · How It Works · Data refreshed daily, snapshot 2026-10-20.