Benchmarks · math
AIME (Vals)
Challenging national math exam given to top high-school students Scores are the values published by VALS AI. Source benchmark version 1.
19 models · 19 results · updated 2026-04-16 · vals.ai ↗
Leaderboard
Muse Spark
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. 10 more models under All models in settings.
More from this benchmark
3 published measuresTime per task96 results · lower is better
9 more models under All models in settings.
Score vs any published measure2 measures: sub-scores, splits, counts
Breakdown
darker is better within each column · click a column to sort| Model | ||
|---|---|---|
| Muse Spark | 96.7 | 97.1 |
| Opus 4.7 max | 96.7 | 95.8 |
| Qwen 3.6 Plus max | 95.4 | 93.8 |
| GLM 4.7 none | 95.0 | 91.7 |
| GLM 4.6 none | 94.6 | 90.8 |
| gpt-oss-120b high | 93.1 | 92.1 |
| Qwen 3.5 Flash max | 93.3 | 91.7 |
| Sonnet 4.6 max | 93.3 | 91.3 |
| GLM 5.1 max | 92.1 | 91.7 |
| MiniMax M2.7 high | 91.7 | 90.4 |
| MiniMax M2.5 high | 89.2 | 88.3 |
| GLM 4.5 xhigh | 87.1 | 86.3 |
| gpt-oss-20b high | 86.7 | 85.4 |
| Qwen 3 Max | 85.8 | 76.3 |
| MiniMax M2.1 | 79.2 | 76.7 |
| DeepSeek R1 xhigh | 77.5 | 70.4 |
| Grok 3 none | 60.4 | 57.1 |
| Grok 4 Fast none | 20.8 | 45.8 |
| DeepSeek V3 none | 30.4 | 24.6 |
All results
19 rows| # | Model | Effort | AIME (Vals) | Subscription $ / task | API $ / task | Time |
|---|---|---|---|---|---|---|
| 1 | Muse Spark | 96.9% | $0.005 | $0.005 | 2.6 min | |
| 2 | Opus 4.7 | max | 96.3% | $0.45 | $0.45 | 3.0 min |
| 3 | Qwen 3.6 Plus | max | 94.6% | $0.046 | $0.094 | 8.6 min |
| 4 | GLM 4.7 | none | 93.3% | $0.043 | $0.043 | 8.8 min |
| 5 | GLM 4.6 | none | 92.7% | $0.12 | $0.12 | 12 min |
| 6 | gpt-oss-120b | high | 92.6% | $0.010 | $0.010 | 2.6 min |
| 7 | Qwen 3.5 Flash | max | 92.5% | $0.011 | $0.011 | 4.0 min |
| 8 | Sonnet 4.6 | max | 92.3% | $0.44 | $0.44 | 5.5 min |
| 9 | GLM 5.1 | max | 91.9% | $0.018 | $0.071 | 6.3 min |
| 10 | MiniMax M2.7 | high | 91.0% | $0.004 | $0.023 | 3.9 min |
| 11 | MiniMax M2.5 | high | 88.8% | $0.021 | $0.041 | 6.2 min |
| 12 | GLM 4.5 | xhigh | 86.7% | $0.047 | $0.047 | 5.6 min |
| 13 | gpt-oss-20b | high | 86.0% | $0.011 | $0.011 | 4.1 min |
| 14 | Qwen 3 Max | max | 81.0% | $0.033 | $0.033 | 17s |
| 15 | MiniMax M2.1 | 77.9% | $0.024 | $0.024 | 4.0 min | |
| 16 | DeepSeek R1 | xhigh | 74.0% | $0.053 | $0.053 | 2.6 min |
| 17 | Grok 3 | none | 58.8% | $0.091 | $0.091 | 64s |
| 18 | Grok 4 Fast | none | 33.3% | $0.001 | $0.001 | 5.9s |
| 19 | DeepSeek V3 | none | 27.5% | $0.003 | $0.003 | 59s |