GLM-5.2 (Baidu)

GLM-5.2 sits within Baidu's evolving GLM family as a text-only large language model positioned around strong logical reasoning. A Baidu-published technical article describes it as an open-source milestone released in 2026 whose value lies in demonstrating dependable reasoning within a pure text architecture, even as that same design left multimodal tasks out of reach. The article frames GLM-5.2 as the foundation on which Baidu's engineering team later layered dynamic attention routing and hierarchical context handling to produce a successor capable of fusing text, images, and code.

For practitioners, GLM-5.2's practical profile centers on text-first workloads where structured reasoning matters more than visual understanding, such as analytical writing, code generation in plain-text contexts, and step-by-step problem solving. Its open-weight availability makes it attractive for teams that want to run or fine-tune a reasoning-oriented model on their own infrastructure rather than relying solely on hosted APIs. Organizations anticipating multimodal needs, however, may find the newer GLM-5.3 a more natural fit, since it preserves GLM-5.2's reasoning strengths while extending the architecture to handle cross-modal input with documented gains in cross-modal question answering accuracy.

LLM Gatewaybaidu/glm-5.2glm

Open provider page Provider docs

Quick Info

Provider: LLM Gateway
Model key: baidu/glm-5.2
Release date: Jun 13, 2026
Last updated: Jun 13, 2026
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $1.40
Output token cost: $4.40

Limits

Output tokens: 131,072 tokens
Context window: 1,048,576 tokens

Transparent token rates

Compare GLM-5.2 (Baidu) pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GLM-5.2 (Baidu)

LLM Gateway

CoverageBenchmark

Morph's "GLM 5: All Versions" overview explicitly names GLM-5.2 as the June 2026 release in Z.ai's GLM-5 family, documenting it as a 753B-parameter mixture-of-experts model with roughly 40B active parameters that expanded context length to 1M tokens while keeping the same base shape as the surrounding GLM-5.x siblings. For a model-focused feed, the page is most useful as family-level context that confirms GLM-5.2 exists as a distinct, named version with a documented release month and the 1M context window milestone, rather than as standalone 5.2 news. The Morph-listed $1.19/$3.74 per-million-token figure for GLM-5.3 is Morph's own bi

LLM Gateway

Coverage

Zhipu AI announced on June 13, 2026 that its GLM-5.2 large language model will be released as open-source software under the MIT license with a 1 million token context window and no regional usage restrictions. The Pandaily report frames the release as a response to tightened US AI export controls, though the core tech On Zhipu's ZCode 3.0 benchmark platform, GLM-5.2 demonstrated practical coding capability by generating, in single HTML files with zero dependencies, a 925-line SVG mechanical clock with five concentric layers and seven interlocking gears, a functional 3D penalty-kick game using Three.js and Cannon.js with an AI goalke

LLM Gateway

CoverageBenchmark

BenchmarkList publishes a model-specific results page for "Z.ai GLM 5.2," listing quantitative eval scores across agentic, tool-use, and autonomous-training benchmarks rather than a launch announcement. Recorded results include a Vending-Bench 2 Final Account Value of $8,313.78 from a $500 starting account across six e As a third-party benchmark aggregator entry rather than a primary news story, the page's value to a model-focused feed is the dated eval evidence for GLM-5.2 specifically, not provider pricing. Notably, every comparative number on the page (e.g., GLM-5.3 at 50.3% Tau3-Banking, Opus 5 at 85.8% MCP Atlas) shows GLM-5.2 t