Llama 3.2 Vision 11B Instruct: benchmark results
Provider: Meta. Released 2024-09-25. Access: Open.
Aggregate Rating 1442 ± 21, rank #1019 of 1640 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| TrustGen - HallusionBench Hard | 52.25 | Accuracy (%) | 61.1 |
| BenchTable - Utility | 56.6 | Weighted Score (%) | 56.5 |
| TrustGen - HallusionBench Overall | 52.74 | Accuracy (%) | 33.3 |
| SnakeBench | 18 | TrueSkill Rating | 31.2 |
| BenchTable - Tech | 19.7 | Weighted Score (%) | 21.5 |
| Wolfram LLM Benchmarking Project | 23.9 | Correct Functionality (%) | 19.2 |
| BenchTable | 18.6 | Total Score (%) | 14.4 |
| TrustGen - HallusionBench Easy | 53.93 | Accuracy (%) | 11.1 |
| Chatbot Arena (Vision - English) | 1017 | Arena Score | 6.7 |
| TrustGen - AutoHallusion Existence | 71.74 | Accuracy (%) | 5.6 |
| Chatbot Arena (Vision) | 993 | Arena Score | 5.2 |
| BenchTable - Reasoning | -1.6 | Weighted Score (%) | 3.1 |
Interactive version: theaggregate.ai/model?slug=llama-3-2-vision-11b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-07.