---
format: "aidr-story-markdown/v1"
id: "4db404cd9fbe7f1c2c5fac1614d39cb1ee5e6aed124147a908fd2893108fead0"
canonical_url: "https://aidr.today/4db404cd?lang=en"
title: "GLM-5.3 Scores 69 on DeepSWE to Rival Fable 5 on Terminal Bench"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-08-20T14:56:35.000Z"
category: "Models"
topics: ["llm","coding","agent","benchmark"]
source_urls: ["https://huggingnews.com/ai/update-glm-53-scores-69-on-deepswe-to-rival-fable-5-on-terminal-bench-8f1a7b7f","https://x.com/ZixuanLi_/status/2090392374077751754","https://x.com/Zai_org/status/2090422632982212779","https://x.com/scaling01/status/2090436220488745457","https://x.com/LiamFedus/status/2090363702042304847","https://x.com/NandoDF/status/2090333885858988299"]
summary: "Z AI’s newest language model showed strong performance in autonomous software engineering during latest test runs. The results on Terminal-Bench 3.0 suggest the GLM-5.3 iteration has reached a capability level nearly equal to Fable 5 for terminal agent tasks, following the release of a model specifically optimized for coding and defensive cybersecurity. GLM-5.3 recorded a score of 69 on the official DeepSWE leaderboard, which measures the ability of AI to solve real-world software issues. Z AI launched the model's API on Aug 18 through partner model gateways, maintaining the same price point as the predecessor GLM-5.2."
---

# GLM\-5\.3 Scores 69 on DeepSWE to Rival Fable 5 on Terminal Bench

> [Open the canonical story](<https://aidr.today/4db404cd?lang=en>)

**Published:** 2026-08-20T14:56:35.000Z
**Category:** Models
**Topics:** llm, coding, agent, benchmark

## Summary

Z AI’s newest language model showed strong performance in autonomous software engineering during latest test runs\. The results on Terminal\-Bench 3\.0 suggest the GLM\-5\.3 iteration has reached a capability level nearly equal to Fable 5 for terminal agent tasks, following the release of a model specifically optimized for coding and defensive cybersecurity\. GLM\-5\.3 recorded a score of 69 on the official DeepSWE leaderboard, which measures the ability of AI to solve real\-world software issues\. Z AI launched the model's API on Aug 18 through partner model gateways, maintaining the same price point as the predecessor GLM\-5\.2\.

## Sources

- [Story source](<https://huggingnews.com/ai/update-glm-53-scores-69-on-deepswe-to-rival-fable-5-on-terminal-bench-8f1a7b7f>)
- [Story source](<https://x.com/ZixuanLi_/status/2090392374077751754>)
- [Supporting source](<https://x.com/Zai_org/status/2090422632982212779>)
- [Supporting source](<https://x.com/scaling01/status/2090436220488745457>)
- [Story source](<https://x.com/LiamFedus/status/2090363702042304847>)
- [Supporting source](<https://x.com/NandoDF/status/2090333885858988299>)