Post-Cutoff
Search For AI agents

Zhipu AI (Z.ai)GLM-5 familyCurrentReasoning modelReleased Aug 2026

GLM-5.3

Z.ai’s top model for long-horizon software engineering, terminal agents and security research; 1M context, 128K output.

Model idglm-5.3

Price$1.40 in, $4.40 out per 1M tokens

Checked13 October 2026

No published training cutoff

At least1,604events since its release in August 2026

72 historic, 356 major. 74 days of news.

Zhipu AI (Z.ai) doesn’t publish a training cutoff, so we count from its release date. The real gap is larger.

Five things it has never heard of

How to call it

9 ways to use GLM-5.3, with the model id each provider expects.

Providers, model ids, endpoints and docs for GLM-5.3
ProviderModel idEndpointDocs
Z.ai API glm-5.3 https://api.z.ai/api/paas/v4 docs.z.ai
Z.ai API (Anthropic format) glm-5.3 https://api.z.ai/api/anthropic docs.z.ai
Alibaba Cloud Model Studio ZHIPU/GLM-5.3 alibabacloud.com
OpenRouter z-ai/glm-5.3 openrouter.ai
Hugging Face huggingface.co
Baseten (GLM-5.3 Fast, speed-optimized) zai-org/GLM-5.3 https://inference.baseten.co/v1 baseten.co
Vercel AI Gateway (GLM 5.3 Fast; routed to Baseten or Alibaba; $2.10 in / $6.60 out) vercel.com
AWS Bedrock (GA Oct 5, 2026; prompt caching supported) aws.amazon.com
Web app chat.z.ai

Pricing

Prices for GLM-5.3
Input$1.40 per 1M tokens
Output$4.40 per 1M tokens
Cache read$0.26 per 1M tokens

As published: “per 1M tokens (USD)”.

Source: docs.z.ai, checked 13 October 2026.

Specs

Input
Text
Output
Text
Context window
1,000,000 tokens
Max output
128,000 tokens
Open weights
Yes
Licence
glm-5.3 (custom)
Training cutoff
Not published
Released
August 2026
Status
Current
Type
Reasoning model

What stands out

What its maker marketed at launch, and what people found later, each with its source.

  • Post-training-only jump in coding

    Same base as GLM-5.2; Z.ai reports +50% on its Code Bench and open-model SOTA on Terminal Bench 3.0 and Agents’ Last Exam (CLI).

    Source: docs.z.ai

  • Emergent cyber capability

    Best CyberGym vulnerability-discovery score to date per Z.ai; exploitation benchmark scores more than double GLM-5.2’s.

    Source: docs.z.ai

  • Always-on reasoning with effort levels

    thinking.type disabled no longer allowed; reasoning_effort low/high/max (default max).

    Source: docs.z.ai

  • Coding Plan integration

    Available in the GLM Coding Plan (points-based; off-peak/weekend calls cost 50% points) for Claude Code, Cline, OpenCode etc.

    Source: docs.z.ai

Notes

Z.ai flagship. GLM 5.3 Fast (speed-optimized GLM-5.3 at $2.10 in / $0.21 cached / $6.60 out, 1M context, 262k max output per Vercel) appeared on Baseten and Vercel AI Gateway around Oct 7, 2026 (LLM Gateway timeline); it is not on Z.ai’s own pricing page as of 2026-10-09 and no Z.ai announcement was found. Text-only input. Migration: requests with thinking disabled fail - set enabled + reasoning_effort low. Coding Plan base URL is https://api.z.ai/api/coding/paas/v4. GLM-5.2 (same price, MIT weights) still listed. GLM-5.3-Highspeed: seen on 2026-10-13 in a gray release inside ZCode 3.15.1 for randomly selected subscribers, about 490 tokens/s in one user’s test (Pandaily citing QbitAI); not announced by Zhipu, no API id or price on docs.z.ai.

curl https://api.z.ai/api/paas/v4/chat/completions \
 -H "Authorization: Bearer $ZAI_API_KEY" -H "Content-Type: application/json" \
 -d '{"model":"glm-5.3","thinking":{"type":"enabled"},"reasoning_effort":"max","messages":[{"role":"user","content":"Hello"}]}'

Sources: https://docs.z.ai/guides/llm/glm-5.3 · https://docs.z.ai/guides/overview/pricing · https://huggingface.co/zai-org/GLM-5.3

Not yet checked against an official source: Release day not verified (OpenRouter 2026-08-18, HF 2026-08-25). A speed-optimized ‘GLM-5.3 Fast’ API appeared on Oct 7, 2026 (LLM Gateway release timeline); Baseten lists it at $2.10 input / $0.21 cached / $6.60 output per 1M tokens, 753B-A40B, custom license; no Z.ai announcement or throughput figure found (unverified).