AstroMind - Intent Inference: leaderboard
Metric: Accuracy (%; 66 maneuver-intent questions, 31 four-option and 35 open-ended answers mapped to the label set by a judge; empty responses excluded). Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 32B | 67.74 |
| 2 | GPT-OSS-20B | 64.62 |
| 3 | Gemma 3 27B | 56.92 |
| 4 | QwQ-32B | 56.25 |
| 5 | Qwen 3 8B | 38.1 |
| 6 | Gemma 3 4B | 25.76 |
Interactive version: theaggregate.ai/benchmark?slug=astromind-intent-inference · How It Works · Data refreshed daily, snapshot 2026-09-26.