Benchmarks · reasoning
HLE-Diamond
HLE-Diamond: the refined 1,000-question version of Humanity's Last Exam from CAIS and Scale AI, split evenly between reasoning and knowledge questions; accuracy.
62 models · 118 results · data as of 2026-11-06 · artificialanalysis.ai ↗
Leaderboard
Gemini 4 Carbon high
Heavy line: best score at each price. Thin lines join one model's effort levels. Cost uses each model's median subscription plan, or API price with no plan data. 27 more models under All models in settings.
More from this benchmark
5 published measuresScore vs any published measure5 measures: sub-scores, splits, counts
47 more models under All models in settings.
Breakdown
darker is better within each column · click a column to sort| Model | |||
|---|---|---|---|
| Gemini 4 Carbon high | 67,031,065 | 1,023,761 | 66,950,265 |
| Opus 5.5 max | 688,331 | 774,834 | 58,907,329 |
| GPT-6 Astra max | 646,278 | 198,139 | 10,696,945 |
| Fable 5.1 max | 684,444 | 591,397 | 55,120,583 |
| GPT-6.1 Sol max | 646,081 | 190,715 | 10,596,437 |
| Gemini 4 Argon high | 17,399,358 | 714,583 | 19,983,320 |
| Sonnet 5.5 max | 639,934 | 527,912 | 79,715,780 |
| GPT-6.1 Terra max | 647,158 | 143,487 | 29,027,705 |
| GPT-6 Sol max | 643,186 | 110,788 | 18,235,715 |
| Opus 5 max | 673,440 | 710,103 | 41,401,351 |
| Muse Wonder (Max) | 559,243 | 158,229 | 41,820,158 |
| Kimi K3.1 max | 730,785 | 150,547 | 36,176,534 |
| Muse Spark 1.3 max | 562,962 | 238,146 | 37,188,219 |
| Gemini 3.8 Flash high | 568,339 | 453,520 | 25,699,399 |
| Hy4 | 531,055 | 190,834 | 43,926,850 |
| Gemini 3.1 Pro | 563,538 | 486,687 | 15,744,051 |
| Haiku 5.5 max | 619,666 | 468,107 | 92,861,341 |
| GPT-5.6 Terra max | 646,471 | 159,152 | 31,510,329 |
| Inkling 2 Small | 511,910 | 4,998,810 | 48,221,875 |
| Kimi K3 max | 727,366 | 170,172 | 23,419,978 |
| Muse Spark Open xhigh | 695,535 | 857,703 | 27,661,361 |
| DeepSeek V4.1 Flash 1017 max | 407,784 | 184,418 | 51,121,738 |
| Grok 4.7 high | 1,832,378 | 175,286 | 44,069,390 |
| GPT-6 Luna max | 642,597 | 140,713 | 29,520,279 |
| Grok 4.6 xhigh | 1,124,775 | 177,999 | 28,403,478 |
All results
118 rows| # | Model | Effort | HLE-Diamond | Subscription $ / task | API $ / task |
|---|---|---|---|---|---|
| 1 | Gemini 4 Carbon | high | 69.1% | $0.58 | $1.63 |
| 2 | Opus 5.5 | max | 67.9% | $0.018 | $1.20 |
| 3 | Gemini 4 Carbon | medium | 67.4% | $0.28 | $0.79 |
| 4 | Opus 5.5 | xhigh | 64.0% | $0.007 | $0.47 |
| 5 | GPT-6 Astra | max | 63.8% | $0.024 | $0.55 |
| 6 | Gemini 4 Carbon | low | 61.9% | $0.24 | $0.69 |
| 7 | GPT-6 Astra | xhigh | 61.5% | $0.016 | $0.37 |
| 8 | Fable 5.1 | max | 60.1% | $0.43 | $2.77 |
| 9 | GPT-6 Astra | high | 59.3% | $0.011 | $0.26 |
| 10 | Fable 5.1 | xhigh | 58.8% | $0.29 | $1.87 |
| 11 | GPT-6.1 Sol | max | 58.8% | $0.006 | $0.11 |
| 12 | Gemini 4 Argon | high | 56.7% | $0.34 | $0.48 |
| 13 | GPT-6.1 Sol | xhigh | 55.8% | $0.004 | $0.066 |
| 14 | Opus 5.5 | high | 55.6% | $0.003 | $0.19 |
| 15 | Sonnet 5.5 | max | 53.3% | $0.012 | $0.80 |
| 16 | GPT-6 Astra | medium | 52.7% | $0.005 | $0.12 |
| 17 | GPT-6.1 Sol | high | 51.9% | $0.003 | $0.043 |
| 18 | Opus 5.5 | medium | 51.4% | $0.002 | $0.13 |
| 19 | Fable 5.1 | high | 50.4% | $0.14 | $0.91 |
| 20 | GPT-6.1 Terra | max | 47.3% | $0.007 | $0.15 |
| 21 | GPT-6.1 Sol | medium | 46.0% | $0.001 | $0.019 |
| 22 | GPT-6 Sol | max | 45.2% | $0.006 | $0.18 |
| 23 | GPT-6 Astra | low | 45.1% | $0.002 | $0.050 |
| 24 | Fable 5.1 | medium | 45.0% | $0.071 | $0.46 |
| 25 | Opus 5 | max | 44.8% | $0.018 | $1.06 |
| 26 | Sonnet 5.5 | xhigh | 44.1% | $0.004 | $0.25 |
| 27 | GPT-6.1 Terra | xhigh | 42.3% | $0.002 | $0.041 |
| 28 | Opus 5.5 | low | 41.7% | $0.001 | $0.035 |
| 29 | Muse Wonder (Max) | max | 39.0% | $0.56 | $0.84 |
| 30 | GPT-6.1 Sol | low | 38.2% | $0.001 | $0.009 |