TutorBench
Chat & Writing · 2025-09-11
Rubric score on tutoring tasks for high-school and AP-level subjects.
Top models (higher is better)
| Model | Score |
|---|---|
| Muse Spark | 68.5 |
| GPT-5.4 Pro | 56.6 |
| Gemini 2.5 Pro Preview 06-05 | 55.6 |
| GPT-5 | 55.3 |
| o3-pro | 54.6 |
| Kimi K2.5 | 54.6 |
| GPT-5.1 | 54.1 |
| Opus 4.6 | 53.7 |
| Gemini 3 Pro Preview | 53.7 |
| GPT-5.2 | 53.5 |
| Gemini 3.1 Pro Preview | 53.0 |
| o3 | 52.8 |
| Gemini 3.1 Flash-Lite Preview | 51.5 |
| Opus 4.5 | 51.2 |
| Opus 4.1 | 50.8 |
| Opus 4 | 49.7 |
| GPT-5.1 Instant | 49.1 |
| Sonnet 4.5 | 49.0 |
| Claude 3.7 Sonnet | 46.5 |
| Llama 4 Maverick Instruct | 40.2 |
| GPT-4o | 36.1 |