OneMillion-Bench: leaderboard
Humanlaya web-scale task benchmark: pass rate % over tasks with an estimated economic value per model; Base and Search-augmented rows are listed separately.
Metric: Pass Rate (%). Source: humanlaya.com. Saturation forecast: Around February 2027. 36 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 36.5 | #67 |
| 2 | GPT-5.4 | 31.5 | #83 |
| 3 | GPT-5.2 | 23 | #115 |
| 4 | Qwen 3.5 Plus | 22.5 | #133 |
| 5 | GPT-5.3 Codex | 18.5 | #72 |
| 6 | Kimi K2.5 | 17.5 | #152 |
| 7 | DeepSeek V3.2 Speciale | 16 | #138 |
| 8 | Gemini 3 Pro (Preview) | 15.5 | #71 |
| 9 | Seed 2.0 Pro | 13.5 | #122 |
| 10 | Grok 4 | 12 | #189 |
| 11 | Qwen 3 Max | 12 | #213 |
| 12 | GLM-5 | 9.5 | #149 |
| 13 | MiniMax-M2.1 | 8 | #299 |
| 14 | Step 3.5 Flash | 7 | #281 |
| 15 | MiniMax-M2.5 | 4 | #316 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=onemillion-bench · How It Works · Data refreshed daily, snapshot 2026-11-05.