Vibe-Bench - Personalized Fine-tuning - Response BERT-F1: leaderboard
Metric: Score (0-100). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen3-8B (Vibe-Bench P-SFT) | 16.53 |
| 2 | Ministral-8B (Vibe-Bench P-SFT) | 15.43 |
| 3 | Llama-3.1-8B (Vibe-Bench P-SFT) | 14.55 |
| 4 | Qwen3-4B (Vibe-Bench P-SFT) | 14.26 |
| 5 | Qwen3-14B (Vibe-Bench P-SFT) | 14.15 |
| 6 | Llama-3.2-3B (Vibe-Bench P-SFT) | 14.13 |
Interactive version: theaggregate.ai/benchmark?slug=vibe-bench-personalized-fine-tuning-response-bert-f1 · How It Works · Data refreshed daily, snapshot 2026-09-24.