GLM-5.3 released: 743B base model, boosts coding and cyber defense
TheZachMueller · x · 2026-08-14
Zhipu released GLM-5.3, built on a 743B base model with post-training for top-tier coding and agentic capabilities, and a major leap in cybersecurity. Benchmarks: Terminal Bench 3.0: 28.3, DeepSWE: 66.9, Agents' Last Exam: 28.5, GDPVal-AA: 1769.
Related event: Zhipu Releases GLM 5.3 with Enhanced Coding and Cybersecurity(20 posts)→
More from Models
- Opus 5 Costs More Per Hour Than a London Developer's Salary, Sparks Debate — kylegawley · 2026-08-14
- Gemini 3.7 Flash vs Opus 5 vs Grok 4.6 vs Kimi K3: Same Sketch, Four Wildly Different Cars — CodeByPoonam · 2026-08-14
- Qwen3.8-27B Model Card Goes Live, Benchmarks Pending — -Cubie- · 2026-08-14
- User tests Gemini on game guide, finds frequent errors, says AI still in infancy — JaxUK89 · 2026-08-14
- Apple reportedly built its own China AI model with Alibaba in rare cross-border deal — The Verge AI · 2026-08-14
- Gemini 3.7 Flash Rolling Out to Web and Mobile Users — testingcatalog · 2026-08-14