Vibe-Bench - BERT Retrieval - Response BLEU-4: leaderboard
Metric: Score (0-100). Source: arxiv.org. Saturation forecast: Around 2029. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemma 3 27B (IT) | 4.34 |
| 2 | Qwen 3 235B A22B 2507 Instruct | 3.84 |
| 3 | Mistral Small 3.2 | 2.93 |
| 4 | Llama 3.1 70B Instruct | 2.79 |
| 5 | Qwen 3 14B | 2.67 |
| 6 | Llama 3.1 8B Instruct | 2.12 |
Interactive version: theaggregate.ai/benchmark?slug=vibe-bench-bert-retrieval-response-bleu-4 · How It Works · Data refreshed daily, snapshot 2026-09-24.