---
title: "GLM 5.2 Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for GLM 5.2."
source: "https://rankedagi.com/models/glm-5-2"
---

# GLM 5.2

| Field | Value |
| --- | --- |
| Organization | Z.ai |
| License | Open Source |
| Type | Thinking |
| Version | Latest |
| Released | 2026-06-16 |
| Context window | 1M |
| Input cost per million tokens | $1.4 |
| Output cost per million tokens | $4.4 |
| Cached input cost per million tokens | $0.26 |
| Last updated | 2026-06-17T19:32:55.052Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 87.7% | RankedAGI Coding Score |  |
| SWEBench Pro | coding | 62.1% | Diverse Agentic Coding |  |
| Terminal-Bench 2.1 | coding | 81.0% | Agentic Terminal Coding | https://www.tbench.ai/leaderboard/terminal-bench/2.1 |
| RankedAGI Agentic | Agents | 82.9% | RankedAGI Agentic Score |  |
| Code Arena | coding | 1595 | Code Arena Elo Score | https://arena.ai/leaderboard/code |
| DeepSWE | coding | 46.2% | Measuring frontier coding agents on original, long-horizon engineering tasks | https://deepswe.datacurve.ai/ |
| RankedAGI Reasoning | reasoning | 80.7% | RankedAGI Reasoning Score |  |
| Humanity's Last Exam | reasoning | 40.5% | Multidisciplinary Reasoning (no tools) |  |
| Humanity's Last Exam w/ Tools | reasoning | 54.7% | Multidisciplinary Reasoning (with tools) |  |
| GPQA Diamond | reasoning | 91.2% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| Text Arena | general | 1471 | ChatArena (LMSYS) ELO Score | https://arena.ai/leaderboard/text |
| AIME 2026 | math | 99.2% | AIME 2026 Competition Math | https://matharena.ai/?comp=aime--aime_2026&view=problem |
| RankedAGI Math | math | 76.3% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 73.4% | Overall RankedAGI score |  |
| MCP Atlas | Agents | 77.0% | Scaled Tool Use |  |
| Code DesignArena | Design | 1360 |  | https://www.designarena.ai/leaderboard/code |
| Toolathlon | Agents | 48.2% | Benchmark to assess language agents' general tool use in realistic environments. | https://toolathlon.xyz/introduction |

## Model Links

| Title | URL |
| --- | --- |
| Blog | https://z.ai/blog/glm-5.2 |
| 𝕏 | https://x.com/Zai_org/status/2066938937344495629 |
| HF | https://huggingface.co/zai-org/GLM-5.2 |