BTZSC - Intent: leaderboard
BTZSC intent slice measures macro-F1 on zero-shot intent classification datasets.
Metric: Macro-F1 (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 35 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen3-Embedding-8B | 58.93 | #685 |
| 2 | Qwen3-Embedding-0.6B | 56.28 | #802 |
| 3 | Qwen 3 8B | 48.15 | #674 |
| 4 | Mistral Nemo Instruct (2407) | 45.51 | #936 |
| 5 | Llama 3.2 3B Instruct | 40.83 | #1352 |
| 6 | Qwen 3 4B | 40.07 | #829 |
| 7 | Phi-4 Mini Instruct | 36.91 | #1245 |
| 8 | Gemma 3 1B (IT) | 23.56 | #1504 |
| 9 | Gemma 3 270M (IT) | 12.96 | #1581 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=btzsc-intent · How It Works · Data refreshed daily, snapshot 2026-10-20.