GLM-4.6V (Reasoning): benchmark results
GLM-4.6V evaluated with reasoning enabled. Provider: Zhipu. Released 2025-12-08. Access: Open.
Aggregate Rating 1580 ± 19, rank #766 of 2168 rated models, from 70 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA AIME 2025 | 85.33 | Accuracy (%) | 83.4 |
| EDCT-Bench (Binary) | 57 | Counterfactual Consistency Score, binary (%; mean over cited | 80 |
| EDCT-Bench (Rubric-Graded) | 59 | Counterfactual Consistency Score, rubric-graded (%; LLM-judg | 80 |
| AA-Omniscience Hallucination Rate | 51.46 | Hallucination Rate (%) | 78.1 |
| UGI - Writing | 39.66 | Writing Score | 70.7 |
| UGI - Natural Intelligence | 28.9 | NatInt Score | 69.8 |
| AA MMLU-Pro | 79.89 | Accuracy (%) | 66.7 |
| AA-Omniscience Index - Health | -26.9 | Omniscience Index | 57 |
| AA-Omniscience Index - Humanities & Social Sciences | -20.6 | Omniscience Index | 56.8 |
| AA-Omniscience Index - Software Engineering (SWE) - Rust | 24 | Omniscience Index | 56.8 |
| AA-Omniscience Index - Law | -32.2 | Omniscience Index | 54.9 |
| AA-Omniscience Index - Business | -24.9 | Omniscience Index | 54.4 |
Interactive version: theaggregate.ai/model?slug=glm-4-6v-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-20.