AI Model Timeline

Tracking the accelerating release frequency of frontier AI models.

GLM-5.3

Zhipu AI·

LLMsopen-weightFrontier

Overview

Open-weight release of Z.ai's GLM-5.3 flagship: a 753B MoE activating ~40B parameters per token with a 1M-token context, architecturally identical to GLM-5.2 — the capability gain is post-training only. Weights landed on Hugging Face on 2026-08-28, fourteen days after launch, once Z.ai's safety review completed — but under a bespoke GLM-5.3 License instead of the MIT license GLM-5.2 shipped under: use, modification, distribution, sublicensing, sale, deployment and fine-tuning are all permitted, but a company with more than $10B aggregate revenue over any 12 consecutive months must pass a Z.ai security review before hosting the model commercially. Individual users and smaller companies are unaffected. The official checkpoint is FP8 (~743 GB of weights) and needs an 8-GPU H100/H200-class node; vLLM and SGLang serve it directly.

Capabilities and innovations

753B total / ~40B active MoE (base unchanged from GLM-5.2)1M-token contextAgentic / long-horizon codingCybersecurity workflowsGLM-5.3 License (custom, revenue-tiered)Capability gain from post-training scaling alone, base model unchangedRevenue-gated open-weight license replacing MIT (security review above $10B revenue)Weights published 14 days after the API launch, after a safety review

Benchmarks

BenchmarkScoreSource
Terminal-Bench 2.x71.54%independent
Terminal-Bench 2.1 71.54%, measured by Vals AI with the Terminus 2 harness (accessed 2026-09-23). Vals AI is the reference harness for this field so values compare across vendors; vendor and other third-party figures stay in the notes below. No Terminal-Bench 2.x value was recorded before.
Terminal-Bench 3.032.4%independent
Terminal-Bench 3.0 32.4% (± 2.9, 95% CI), official leaderboard maintained by the Terminal-Bench team (Stanford / Harbor / Laude Institute), max effort, agent harness: Claude Code (accessed 2026-09-23). Mirrored from the cloud entry (zhipu-3). Z.ai self-reported 28.3 in the GLM-5.3 launch table; the leaderboard value is kept as primary per the vendor-vs-independent rule. Do not compare against the terminal_bench column: 3.0 is a harder generation on a different scale.
Terminal-Bench 4.041.8%independent
Terminal-Bench 4.0 41.8% (± 3.2, 95% CI), official leaderboard maintained by the Terminal-Bench team (Stanford / Harbor / Laude Institute), max effort, agent harness: Claude Code (accessed 2026-09-23). 4.0 is its own generation and not comparable with the 2.x or 3.0 fields.
DeepSWE v1.166.9%vendor
DeepSWE v1.1 66.9 from Z.ai's GLM-5.3 launch table at maximum thinking effort, reported in place of SWE-bench Pro (GLM-5.2: 46.2 in the same table). Vendor self-report over its own model. Datacurve's own DeepSWE board was not reachable from this environment, so the value is not cross-checked against it.
  • GPQA Diamond: independent evaluation pending — Z.ai published no GPQA table for this release; Vals AI and Artificial Analysis had no GLM-5.3 entry as of 2026-08-31.
  • Humanity’s Last Exam (no tools): not reported by the vendor — Z.ai reports HLE with tools only, which is not comparable to the no-tools HLE used here (same handling as the GLM-5.2 entry).
  • SWE-bench Pro: not reported by the vendor — Z.ai reports DeepSWE v1.1 (66.9, GLM-5.2: 46.2, both now recorded in deepswe_v1_1) instead of SWE-bench Pro.
  • MMLU-Pro: independent evaluation pending — Not in Z.ai's table; Vals AI had no GLM-5.3 entry as of 2026-08-31.
  • MMMU-Pro: not reported by the vendor

Architecture and hardware

Parameters
753B total, 40B active per token (MoE)
Estimated VRAM at Q4
~433 GB, Frontier class
Quantization formats
FP8
Recommended runtime
vLLM
License
GLM-5.3 License (custom, revenue-tiered)

Links

More from Zhipu AI

Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.