JMeetEval - Instruction Level: leaderboard
Metric: Instruction-level strict accuracy (%). Source: github.com. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | gemma-4-E4B-it | 83.3 |
| 2 | gemma-4-E2B-it | 80.6 |
| 3 | Qwen 3 VL 8B (Thinking) | 80.6 |
| 4 | Qwen 3 8B | 76 |
| 5 | Qwen 3.5 4B | 72.1 |
Interactive version: theaggregate.ai/benchmark?slug=jmeeteval-instruction-level · How It Works · Data refreshed daily, snapshot 2026-09-05.