CoS-VQ BF16 - LAMBADA OpenAI: leaderboard
The CoS-VQ study reports these reference results for unquantized models on language modeling and zero-shot questions. Lower perplexity is better. Zero-shot accuracy (%).
Metric: Zero-shot accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 70B Instruct | 75.68 | #559 |
| 2 | Llama 3.1 8B Instruct | 73.08 | #1035 |
| 3 | Qwen 3 14B | 67.88 | #536 |
| 4 | Qwen 3 32B | 67.18 | #443 |
| 5 | Qwen 3 8B | 64.08 | #671 |
| 6 | Qwen 3 4B | 59.4 | #833 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=cos-vq-bf16-lambada-openai · How It Works · Data refreshed daily, snapshot 2026-10-29.