GLM-4V-9B: benchmark results
Provider: Zhipu. Released 2024-06-05. Access: Open.
Aggregate Rating 1457 ± 16, rank #3458 of 10916 rated models, from 572 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MUSE - Jigsaw Puzzle | 42 | Accuracy (%) | 100 |
| OpenVLM A-Bench - Contradiction Overcome | 95.8 | Accuracy (%) | 99.4 |
| OpenVLM ScienceQA Val | 98.7 | Accuracy (%) | 99.1 |
| OpenVLM MMT-Bench - Attribute Hallucination | 90 | Score (%) | 98.8 |
| OpenVLM VCR - Chinese Easy | 53.4 | Jaccard (%) | 95.9 |
| OpenVLM VCR - Chinese Hard | 15.3 | Jaccard (%) | 95.9 |
| OpenVLM MMT-Bench - National Flag Recognition | 100 | Score (%) | 94.9 |
| OpenVLM MMT-Bench - Pixel Recognition | 75 | Score (%) | 94.7 |
| OpenVLM Q-Bench - How (Others) | 77.2 | Accuracy (%) | 94 |
| OpenVLM MMT-Bench - Temporal Localization | 63.2 | Score (%) | 93.7 |
| OpenVLM VCR | 49.8 | Overall Jaccard (%) | 93.6 |
| OpenVLM MMT-Bench - Astronomical Recognition | 88.9 | Score (%) | 93 |
Interactive version: theaggregate.ai/model?slug=glm-4v-9b · How It Works · Data refreshed daily, snapshot 2026-11-08.