GLM Timeline: Every Zhipu Model Update

Doubao Work team and enterprise customers can now use third-party models without buying from each model vendor, paying with their existing AI points. (Chinese-language source)

A paper from Tsinghua and Zhipu reveals GLM-Omni, an internal unified multimodal model, and finds that knowledge it learns through image generation rarely carries over to text answers.

NVIDIA is promoting local inference, but every setup it has validated for its FP4 GLM-5.3-Flash checkpoint uses two GPUs or two systems.

By rewriting an agent's harness at test time, one refinement step raised GLM-5 from 64.2 to 75.7 on WorkBuddy-Bench.

Chinese-language source: QbitAI reports that GLM-Legal, built on GLM-5.3-Flash, ranks first among open-weight models on both Harvey LAB and LegalBench.

A new Zhipu model is rolling out at random inside Zcode, and a user reports output of about 490 tokens/s. At that speed, generating 10,000 tokens takes about 20 seconds (Chinese-language source).

Goodfire reads a model's internal computations with small probes to flag risky agent behavior, instead of paying a second AI to reread every step. The product is now live on Baseten.

Z.ai's GLM-5.3 went live on Amazon Bedrock on October 5th with AWS-managed inference, but only eligible enterprise customers can use it.

Two trained checkpoints produced tool-call failure rates of 2.24% and 1.77% in a live A/B test reported by Perplexity.

Days after a snapshot-upload controversy, Zhipu released every ZCode component under Apache 2.0 and cut the local repository snapshot pipeline.

Spoken question in, spoken answer out — no transcription step, no separate text LLM, and up to 77.1% on GSM8K.

Taiyuan Chengming Technology wants answers on deletion, data flow, operation logs and possible cross-border transfer by October 10; Zhipu has not responded publicly.

Zhipu writes up its self-improvement progress and cites one inference-optimization case on a domestic-chip cluster.

The company apologized, says the issue is fixed, and will open-source the codebase and bring in third-party review.

Two weeks from first run to carrying all production traffic, with end-to-end throughput up 3.2x. Chinese-language source.

The raise combines roughly $2 billion in share placement with roughly $3 billion in convertible bonds, earmarked for the next GLM base model, a self-training system and compute infrastructure.

Zhipu's first interim report since listing shows revenue shifting from on-premises deployment to its open platform, with ARR at $1.6B at the end of August. (Chinese-language source)

vLLM wires HiSparse in as a pressure-driven memory tier, so requests keep decoding when their KV cache no longer fits in GPU memory.

A startup called Abliteration.ai has turned guardrail removal into a hosted service, and its lineup includes Z.ai's recently released GLM-5.3.

Zhipu opens a nightly window for paying GLM Coding Plan users, 23:00 to 09:00, with terms that depend on the client. (Chinese-language source)

Zhipu AI opened a Tmall flagship store selling GLM Coding Plan token subscriptions, reportedly the first listed large-model company to put its core AI product on a mainstream e-commerce shelf.

On the coding plan's first anniversary, Z.ai hands every current subscriber a card that resets both the five-hour and weekly limits.

Tencent's OS-level assistant Marvis added custom models on September 1, letting users connect third-party models or run a local one.

Z.ai releases downloadable weights for GLM-5.3, its flagship agentic coding and cyber-defense model previously offered only through hosted products.

Two Chinese labs shipped open-weight models a day apart, designed independently, yet their configs read like near-copies.

Zhipu confirmed the mysterious "Ox Alpha" model is GLM-5.3-Flash; during its anonymous test it briefly became OpenRouter's most-used model of the week, running entirely on a domestic-chip cluster.

A code comparison found that GLM-5.3-Flash and Kimi K3 both use KDA linear attention with a lower forget-gate bound of -5.

Z.ai has opened the first natively multimodal GLM-5 model, a 320B-total, 18B-active MoE.

Z.AI confirmed that the anonymous Ox Alpha model is a new GLM iteration and planned to release its weights on August 26.

Z.ai put its coding/cybersecurity model behind a hosted API on Aug 18, keeping GLM-5.2 pricing and a 1M-token context while it holds back the weights.

Z.ai released GLM-5.3 on August 14, reusing the 743B GLM-5.2 base unchanged, with every gain coming from more post-training.

Writer post-trains the open GLM-5.2, claiming up to 50% lower cost on basic tasks.

The figure is self-reported and does not separate registered, active or paying users; no revenue or retention data.

An open-weight Chinese model trails top labs by only months on cyber and bio, with a widening safety gap.

Per Pandaily, Zhipu is building a 1GW all-domestic-chip data center, acquiring compiler startup Zhongke Jiahe, and exploring custom AI chips.