AlpacaEval 1.0: leaderboard
Original instruction-following evaluation using raw win rates against text-davinci-003. 805 prompts scored by an LLM judge. Predecessor to AlpacaEval 2.0.
Metric: Win Rate (%). Source: tatsu-lab.github.io. Saturation forecast: Estimated already saturated. 102 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Mistral Medium | 96.83 | #792 |
| 2 | Xwin-LM-70B-V0.1 | 95.57 | #982 |
| 3 | GPT-4 | 95.28 | #438 |
| 4 | tulu-2-dpo-70B | 95.03 | #1052 |
| 5 | Mixtral 8x7B Instruct (v0.1) | 94.78 | #1073 |
| 6 | GPT-4 (0314) | 94.78 | #574 |
| 7 | Yi Chat 34B | 94.08 | #976 |
| 8 | GPT-4 (0613) | 93.78 | #491 |
| 9 | Mistral 7B Instruct (v0.2) | 92.78 | #1333 |
| 10 | Llama 2 70B Chat (HF) | 92.66 | #1264 |
| 11 | LMCocktail-10.7B-v1 | 92.22 | #1008 |
| 12 | Claude 2 | 91.36 | #761 |
| 13 | zephyr-7B-beta | 90.6 | #1330 |
| 14 | WizardLM-13B-V1.2 | 89.17 | #1425 |
| 15 | vicuna-33B-v1.3 | 88.99 | #1347 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=alpacaeval-1-0 · How It Works · Data refreshed daily, snapshot 2026-10-20.