Up to four models
GLM-5.2 vs Grok Build 0.1
On the index (GPT-4 = 100), GLM-5.2 scores 154 and Grok Build 0.1 141. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
154138–169
WeightsOpen
ReleasedJun 2026
Price in / out$0.0070 / $6.8
Context1.05M
Core tests taken6 of 9
Grok Build 0.1
141114–169
WeightsClosed
ReleasedMay 2026
Price in / out$1.0 / $2.0
Context256K
Core tests taken0 of 9
GLM-5.2 and Grok Build 0.1 are too close to call (12 points apart, within the uncertainty), at 1.4× the price.
Test by testthird-party results
Every result as a table
| Test | GLM-5.2 | Grok Build 0.1 |
|---|---|---|
| Reasoning & knowledge | ||
| LiveBench Data Analysis | 73.7% | 70.8% |
| LiveBench IF | 62.3% | 65.2% |
| LiveBench Language | 76.2% | 72.5% |
| LiveBench Reasoning | 78.6% | 76.4% |
| Maths | ||
| LiveBench Mathematics | 89.8% | 78.4% |
| Coding | ||
| LiveBench Agentic Coding | 51.8% | 45.8% |
| LiveBench Coding | 79.7% | 65.4% |
Tests at least two of these models have taken. Bold rows are the core tests.