GLM-4.1V-9B (Thinking): benchmark results
GLM-4.1V-9B evaluated with thinking enabled. Provider: Zhipu. Released 2025-07-01. Access: Open.
Aggregate Rating 1543 ± 5, rank #908 of 2168 rated models, from 859 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DatBench - Chart Understanding | 76 | Accuracy (%) | 100 |
| M3 Hate Memes (Meme + Post) - Category Macro-F1 | 64.49 | Macro-F1 (%) over the eight hate categories (religion, polit | 100 |
| M3 Hate Memes (Meme + Post) - Category Subset Accuracy | 33.76 | Subset accuracy (%): share of M3's 1,318 hate memes whose pr | 100 |
| M3 Hate Memes (Meme Only) - Category Macro-F1 | 69.02 | Macro-F1 (%) over the eight hate categories (religion, polit | 100 |
| MEGA-Bench Task - 3D Indoor Scene Text BBox Selection | 71.4 | Task Score (%) | 100 |
| MEGA-Bench Task - Booking Web Recommendation | 80.9 | Task Score (%) | 100 |
| MEGA-Bench Task - CLEVRER Video Moving Object Property Recognition | 100 | Task Score (%) | 100 |
| MEGA-Bench Task - Electricity Plot Future Prediction | 92.4 | Task Score (%) | 100 |
| MEGA-Bench Task - Electrocardiogram | 45.7 | Task Score (%) | 100 |
| MEGA-Bench Task - IconQA Count And Reasoning | 68.4 | Task Score (%) | 100 |
| MEGA-Bench Task - Multiview Reasoning Camera Moving | 78.6 | Task Score (%) | 100 |
| MEGA-Bench Task - NExT-QA OE | 60 | Task Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=glm-4-1v-9b-thinking · How It Works · Data refreshed daily, snapshot 2026-10-20.