Back to BenchmarksFrontier
ANLI
Adversarial NLI - challenging natural language inference
Leaderboard
(9 models)| Rank | Model | Score | Stderr |
|---|---|---|---|
| 1 | Phi-3-small-8k-instruct | 58.10 | — |
| 2 | gpt-3.5-turbo-1106 | 58.10 | — |
| 3 | Meta-Llama-3-8B-Instruct | 57.30 | — |
| 4 | Phi-3-medium-128k-instruct | 55.80 | — |
| 5 | Mixtral-8x7B-v0.1 | 55.20 | — |
| 6 | Phi-3-mini-4k-instruct | 52.80 | — |
| 7 | gemma-7b | 48.70 | — |
| 8 | Mistral-7B-v0.1 | 47.10 | — |
| 9 | Phi-4 | 42.50 | — |
Data source: Epoch AI, “Data on AI Benchmarking”. Published at epoch.ai
Licensed under CC-BY 4.0