SAGE - Calculus
Math · 2025-10-08
The Calculus split of SAGE. This child benchmark separates a source-reported subtask or subtrack from the parent aggregate so scores at different grains do not share one benchmark_slug.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 4 Argon | 84.9 |
| Gemini 3.5 Flash | 84.0 |
| Claude Fable 5.1 | 83.6 |
| Claude Opus 5.5 | 83.6 |
| Gemini 3.7 Flash | 83.6 |
| Claude Fable 5 | 82.4 |
| GPT-6.1 Sol | 79.9 |
| GPT-6 Astra | 79.9 |
| Kimi K3 | 79.9 |
| Gemini 3.8 Flash | 79.6 |
| Claude Sonnet 5.5 | 78.4 |
| Gemini 3.1 Pro Preview | 78.4 |
| Claude Opus 5 | 77.2 |
| Opus 4.7 | 75.9 |
| Opus 4.8 | 75.9 |
| Gemini 3.6 Flash | 75.3 |
| GPT-6 Sol | 75.0 |
| Muse Spark 1.1 | 74.7 |
| Opus 4.5 | 73.5 |
| Opus 4.6 | 72.5 |
| GPT-5.5 | 72.2 |
| Qwen3.8-Max | 71.9 |
| Claude Haiku 5.5 | 71.6 |
| Gemini 3 Flash Preview | 71.0 |
| GPT-5.4 | 71.0 |
| Qwen3.7-Plus (source unspecified) | 71.0 |
| Gemini 2.5 Pro | 70.7 |
| Muse Spark 1.2 | 70.7 |
| Grok 4.7 | 70.4 |
| GPT-5.1 | 69.8 |
| Gemini 3 Pro Preview | 69.4 |
| GPT-5.6 Sol | 67.3 |
| Sonnet 5 | 64.5 |
| GPT-5.6 Luna | 64.5 |
| GPT-5 | 63.0 |
| GPT-5.2 | 62.0 |
| Grok 4.6 | 61.7 |
| Sonnet 4.5 | 60.8 |
| Sonnet 4.6 | 60.5 |
| Gemini 3.5 Flash-Lite | 60.5 |