Vector Eval - HellaSwag: leaderboard
Vector Institute Inspect Evals run of HellaSwag, testing commonsense event completion by choosing plausible continuations for everyday scenarios.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 92.43 | #343 |
| 2 | Claude 3.5 Sonnet | 92.28 | #350 |
| 3 | Gemini 1.5 Pro | 91.24 | #445 |
| 4 | Mistral Large 2 (Jul) | 90.48 | #484 |
| 5 | GPT-4o Mini | 87.5 | #599 |
| 6 | Llama 3.1 70B Instruct | 86.99 | #561 |
| 7 | Gemini 1.5 Flash | 85.57 | #618 |
| 8 | Qwen3 Coder Next | 84.33 | #326 |
| 9 | Command-R+ | 79.55 | #891 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=vector-eval-hellaswag · How It Works · Data refreshed daily, snapshot 2026-10-20.