CulMind-R - ReaScore: leaderboard
Metric: ReaScore (0-100): task-adaptive score of the reasoning process, each task's relevant dimensions weighted and scored against expert reasoning, on the 24-task CulMind-R reasoning subset, the model writing a structured reasoning process before its answer; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 64.7 |
| 2 | GPT-5.5 | 59.4 |
| 3 | Qwen 3.5 27B | 58.5 |
| 4 | Qwen 3 VL 8B Instruct | 58.4 |
| 5 | Qwen 3.5 9B | 57.7 |
| 6 | Qwen 3 VL 32B Instruct | 57.2 |
| 7 | GLM-4.5V | 55.2 |
| 8 | Gemini 2.5 Flash | 55.1 |
| 9 | GPT-5.4 Mini | 53.3 |
| 10 | InternVL3-8B | 48.2 |
| 11 | InternVL3-14B | 48 |
Interactive version: theaggregate.ai/benchmark?slug=culmind-r-reascore · How It Works · Data refreshed daily, snapshot 2026-09-29.