Vector Eval - ARC-E: leaderboard
Vector Institute Inspect Evals run of ARC Easy, testing grade-school science knowledge and commonsense reasoning with accuracy scoring.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 99.16 | #343 |
| 2 | Claude 3.5 Sonnet | 99.16 | #350 |
| 3 | O1 | 99.12 | #183 |
| 4 | Qwen3 Coder Next | 98.82 | #326 |
| 5 | Gemini 1.5 Pro | 98.78 | #445 |
| 6 | DeepSeek R1 | 98.74 | #251 |
| 7 | Mistral Large 2 (Jul) | 98.53 | #484 |
| 8 | Gemini 1.5 Flash | 98.4 | #618 |
| 9 | Llama 3.1 70B Instruct | 98.23 | #561 |
| 10 | GPT-4o Mini | 97.94 | #599 |
| 11 | O3 Mini | 97.56 | #272 |
| 12 | Command-R+ | 93.77 | #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-arc-e · How It Works · Data refreshed daily, snapshot 2026-10-20.