GLM-5.2vsGLM-4.7
Across 7 shared benchmarks, GLM-5.2 leads overall: GLM-5.2 wins 7, GLM-4.7 wins 0, with 0 ties and an average score difference of +13.40.
GLM-5.27 wins(100%)(0%)0 winsGLM-4.7
Benchmark scores
Grouped by capability, sorted by largest gap within each. 7 shared benchmarks.
General Knowledge
GLM-5.2 2/2| Benchmark | GLM-5.2 | GLM-4.7 | Diff |
|---|---|---|---|
| LiveBench | 76.249 / 115Normal (No Tools) | 58.0978 / 115Normal (No Tools) | +18.15 |
| HLE | 54.7015 / 181Thinking (With Tools) | 42.8055 / 181 | +11.90 |
AI Agent - Tool Usage
GLM-5.2 1/1| Benchmark | GLM-5.2 | GLM-4.7 | Diff |
|---|---|---|---|
| MCP-Atlas | 76.8013 / 38Thinking (With Tools) | 58.1033 / 38Normal (With Tools) | +18.70 |
Coding and Software Engineer
GLM-5.2 1/1| Benchmark | GLM-5.2 | GLM-4.7 | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 62.109 / 57Thinking (With Tools) | 40.6053 / 57 | +21.50 |
Commonsense Reasoning
GLM-5.2 1/1| Benchmark | GLM-5.2 | GLM-4.7 | Diff |
|---|---|---|---|
| SimpleBench | 58.8020 / 67Normal (No Tools) | 47.7033 / 67Thinking (No Tools) | +11.10 |
General Evaluation
GLM-5.2 1/1| Benchmark | GLM-5.2 | GLM-4.7 | Diff |
|---|---|---|---|
| GPQA Diamond | 91.8626 / 226最高（无工具） | 85.7081 / 226 | +6.16 |
Math and Reasoning
GLM-5.2 1/1| Benchmark | GLM-5.2 | GLM-4.7 | Diff |
|---|---|---|---|
| AIME 2026 | 99.201 / 19Thinking (No Tools) | 92.909 / 19 | +6.30 |
Specs
| Field | GLM-5.2 | GLM-4.7 |
|---|---|---|
| Publisher | 智谱AI | 智谱AI |
| Release date | 2026-06-13 | 2025-12-22 |
| Model type | Reasoning model | Chat model |
| Architecture | MoE | MoE |
| Parameters | 753.33B | 358B |
| Context length | 1M | 200K |
| Max output | 128K | 132072 |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | GLM-5.2 | GLM-4.7 |
|---|---|---|
| Text input | $1.4 / 1M tokens | ¥4 / 1M tokens |
| Text output | $4.4 / 1M tokens | ¥16 / 1M tokens |
| Cache read | $0.26 / 1M tokens | ¥2 / 1M tokens |
| Cache write | Not public | ¥0 / 1M tokens |
Summary
- GLM-5.2leads in:General Knowledge (2/2), AI Agent - Tool Usage (1/1), Coding and Software Engineer (1/1), Commonsense Reasoning (1/1), General Evaluation (1/1), Math and Reasoning (1/1)
On average across the 7 shared benchmarks, GLM-5.2 scores 13.40 higher.
Largest single-benchmark gap: SWE-Bench Pro - Public — GLM-5.2 62.10 vs GLM-4.7 40.60 (+21.50).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.