Skip to content
AI Atlas
ModelActiveOpen weightsIdentity probable

Z.ai GLM 5.2

Z.ai (Zhipu AI)family · GLM5.2docs.mistral.ai/models/zai-glm-5-2

A third-party open source text model from Z.ai with a 1M-token context window.

Open in Graph
data quality76

Updated 3 h ago · first seen 11 Sept 2026

model_01M2943ZSHC4RGJCVFH2K0A6B2

Overview

Identity

Canonical model
Yesidentity confidence: mediumOne row per real model release. Artifacts (checkpoints, quantisations, conversions) and folded evaluation variants point here.
Official checkpoints
official_checkpoints = hf_repo identifiers carried by the model itself; artifacts are separate entities pointing here through canonical_id.
Artifacts
2 quantizations0 official · 0 third-partySeparate entities (checkpoint · quantization · conversion · packaging) pointing to this model through canonical_id.
Provider deployments
3
API aliases
glm-5-2glm-5-2-non-reasoningz-ai/glm-5.2Identifiers under which providers and evaluators refer to this model.
Folded evaluation variants
1Effort / thinking variants (…-high, …-non-reasoning) are result configurations of this model, not separate models. Their old URLs redirect here.

Openness

Open weights— weights downloadable under MIT; commercial use allowed; redistribution allowed; derivatives allowed; 4 dimensions unknown.

Weights downloadable under a permissive or Creative Commons licence allowing commercial use; code or data may be missing.

  • Weights

    Yes

  • Inference code

    —

  • Training code

    —

  • Training data

    —

  • Dataset

    —

  • Commercial use

    Yes

  • Redistribution

    Yes

  • Derivatives

    Yes

Licence: MIT License (permissive · SPDX MIT · stated as “mit”)

dimensions marked null are unknown, not false

Key facts

Release date

Source:OpenRouter public model & pricing listingT2observed 13 h agomedium

Version

Source:Mistral AI docsT1observed 14 h agohigh

Official page
docs.mistral.ai/models/zai-glm-5-2

Source:Mistral AI docsT1observed 14 h agohigh

Model card
huggingface.co/zai-org/GLM-5.2

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 13 h agomedium

Openrouter id

Source:OpenRouter public model & pricing listingT2observed 13 h agomedium

Architecture

Architecture

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

Model type

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

Parameters

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 13 h agomedium

Weights dtype

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

File size

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

Library name

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

Pipeline tag

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 13 h agomedium

Hugging Face repo

Source:OpenRouter public model & pricing listingT2observed 13 h agomedium

Capabilities

Modalities

Modalities
text
Input
text
Output
text

Capabilities

  • Tool calling

    Yes

    Mistral AI docs · T1

  • Structured output

    Yes

    Mistral AI docs · T1

  • Reasoning

    Yes

    OpenRouter public model & pricing listing · T2

  • Vision

    Unavailable

  • Audio

    Unavailable

  • Fine-tuning available

    Unavailable

Context window

Source:OpenRouter public model & pricing listingT2observed 3 h agomedium

Max output

Source:OpenRouter public model & pricing listingT2observed 13 h agomedium

Languages

Source:Hugging Face Hub (public pages, model cards, papers)T2observed 12 h agomedium

Benchmarks32

Compare with another model →

Comparable same task and conditions · Partially comparable same task, conditions differ (effort, temperature, judge) · Not comparable different variant or metric

Benchmark results grouped by comparability group
Benchmark · groupBest scoreTrustConfigurationResultsvs leaderEvaluatedSource
Terminal-Benchagentic · accuracy · variant=hard · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisvarianthardreasoning_effortmaxconditions differ across rows → partially comparable2−15.1 ptvs gpt-5.6-solobs. 12 Sept 2026artificialanalysis.aiT2
Terminal-Benchagentic · accuracy · variant=v2.1 · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisvariantv2.1reasoning_effortmaxconditions differ across rows → partially comparable4non-primary groupobs. 12 Sept 2026artificialanalysis.aiT2
τ²-benchagentic · pass^1 · variant=Telecom · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantTelecomreasoning_effortmaxconditions differ across rows → partially comparable1non-primary groupobs. 12 Sept 2026artificialanalysis.aiT2
τ²-benchagentic · pass^1 · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysis1current leaderobs. 11 Sept 2026artificialanalysis.aiT2
LiveBench Agentic Codingcoding · average score · variant=Agentic CodingOfficial boardvariantAgentic Codingrelease2026-06-251−25.5 ptvs DeepSeek-V4.1-Flash25 Jun 2026livebench.aiT2
LiveBench Codingcoding · average score · variant=CodingOfficial boardvariantCodingrelease2026-06-251−6.72 ptvs Claude Fable 5.125 Jun 2026livebench.aiT2
SciCodecoding · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoning_effortmaxconditions differ across rows → partially comparable2−11.9 ptvs Claude Fable 5.1obs. 12 Sept 2026artificialanalysis.aiT2
Artificial Analysis Intelligence Indexcomposite · indexIndependentreasoning_effortmaxversion4.3conditions differ across rows → partially comparable4−14.7vs Claude Fable 5.1obs. 12 Sept 2026artificialanalysis.aiT2
LiveBenchgeneral · global_averageOfficial boardrelease2026-06-251−10.3 ptvs Claude Fable 5.125 Jun 2026livebench.aiT2
LiveBench Data Analysisgeneral · average score · variant=Data AnalysisOfficial boardvariantData Analysisrelease2026-06-251−9.23 ptvs gpt-6-astra25 Jun 2026livebench.aiT2
LiveBench Languagegeneral · average score · variant=LanguageOfficial boardvariantLanguagerelease2026-06-251−14.4 ptvs Claude Fable 525 Jun 2026livebench.aiT2
IFBenchinstruction-following · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoning_effortmaxconditions differ across rows → partially comparable2−10.0 ptvs Grok 4.3obs. 12 Sept 2026artificialanalysis.aiT2
LiveBench Instruction Followinginstruction-following · average score · variant=IFOfficial boardvariantIFrelease2026-06-251−19.1 ptvs Gemini 3.8 Flash25 Jun 2026livebench.aiT2
Humanity's Last Examknowledge · accuracy · evaluator=Artificial AnalysisIndependentevaluatorArtificial Analysisreasoning_effortmaxconditions differ across rows → partially comparable4−18.0 ptvs Claude Fable 5.1obs. 12 Sept 2026artificialanalysis.aiT2
LiveBench Mathematicsmath · average score · variant=MathematicsOfficial boardvariantMathematicsrelease2026-06-251−7.23 ptvs Claude Fable 5.125 Jun 2026livebench.aiT2
GPQA Diamondreasoning · accuracy · variant=Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantDiamondreasoning_effortmaxconditions differ across rows → partially comparable2non-primary groupobs. 12 Sept 2026artificialanalysis.aiT2
GPQA Diamondreasoning · accuracy · variant=GPQA Diamond · evaluator=Artificial AnalysisIndependentevaluatorArtificial AnalysisvariantGPQA Diamond2−6.77 ptvs gpt-6-astraobs. 11 Sept 2026artificialanalysis.aiT2
LiveBench Reasoningreasoning · average score · variant=ReasoningOfficial boardvariantReasoningrelease2026-06-251−14.0 ptvs gpt-6-astra25 Jun 2026livebench.aiT2

Current rows only, grouped by benchmark → canonical metric → comparability group (task configuration). Effort variants folded into this model appear as rows of the same group. 32 current rows in total. “vs leader” compares with the current leader of the benchmark's primary group only; other groups are not directly comparable. Comparability rules →

Providers & Pricing5

All offers in the price terminal →
Provider deployments, cheapest output first
ProviderContextInput / 1MCached inOutput / 1MNative unitsStatusObservedSource
OpenRoutercheapest outputz-ai/glm-5.2202.8Kout 182.5K$0.15—active4 min agosince 12 Sept 2026openrouter.aiT2
Z.ai APIz-ai/glm-5.2202.8Kout 182.5K$0.15—active4 h agosince 11 Sept 2026openrouter.aiT2
OpenRouterz-ai/glm-5.2:batch1.05Mout 943.7K$0.07—active4 min agosince 12 Sept 2026openrouter.aiT2
Z.ai APIz-ai/glm-5.2:batch1.05Mout 943.7K$0.07—active4 h agosince 11 Sept 2026openrouter.aiT2
Mistral AI La Plateforme——eur_input_per_mtok=1.19eur_output_per_mtok=3.74active14 h agosince 11 Sept 2026docs.mistral.aiT1

USD per 1M tokens as published by each provider; native units (per-request fees, flex/priority tiers) are kept verbatim. Rows are append-only — every price change is kept in the history below. Cost of a workload →

Price history

Step lines per provider; amber markers are recorded changes. Click a marker or a row for the evidence behind that price.

Output price · USD / 1M tokens 3 providers

Output price history of Z.ai GLM 5.2$0$2$4$6Sept 26Sept 26Sept 26Sept 26Sept 26Mistral AI La Plateforme: first observed → $4.4 · 11 Sept 2026Z.ai API: first observed → $2 · 11 Sept 2026Z.ai API: $2 → $2.2 · 11 Sept 2026OpenRouter: first observed → $2 · 12 Sept 2026OpenRouter: $2 → $2.2 · 12 Sept 2026
  • Mistral AI La Plateforme
  • Z.ai API
  • OpenRouter
  • OpenRouter$2 → $2.212 Sept 2026
  • OpenRouterfirst observed $212 Sept 2026
  • Z.ai API$2 → $2.211 Sept 2026
  • Z.ai APIfirst observed $211 Sept 2026
  • Mistral AI La Plateformefirst observed $4.411 Sept 2026

Input price · USD / 1M tokens 3 providers

Input price history of Z.ai GLM 5.2$0$0.50$1$1.5$2Sept 26Sept 26Sept 26Sept 26Sept 26Mistral AI La Plateforme: first observed → $1.4 · 11 Sept 2026Z.ai API: first observed → $0.60 · 11 Sept 2026Z.ai API: $0.60 → $0.70 · 11 Sept 2026OpenRouter: first observed → $0.60 · 12 Sept 2026OpenRouter: $0.60 → $0.70 · 12 Sept 2026
  • Mistral AI La Plateforme
  • Z.ai API
  • OpenRouter
  • OpenRouter$0.60 → $0.7012 Sept 2026
  • OpenRouterfirst observed $0.6012 Sept 2026
  • Z.ai API$0.60 → $0.7011 Sept 2026
  • Z.ai APIfirst observed $0.6011 Sept 2026
  • Mistral AI La Plateformefirst observed $1.411 Sept 2026

Hardware fit37

Estimated

3 of 37 device × quantization combinations fit.

Run locally: your machine →
Estimated hardware fit
HardwareQuantizationDevice memoryEst. memoryFits
Apple M3 Ultra4bit—433.7 GB est.Yes
Mac Studio (Apple M5 Ultra)4bit—433.7 GB est.Yes
NVIDIA DGX B2004bit1,440 GB433.7 GB est.Yes
Apple M2 Ultra4bit—433.7 GB est.No
Apple M1 Ultra4bit—433.7 GB est.No
Apple M3 Max4bit—433.7 GB est.No
Apple M4 Max4bit—433.7 GB est.No
Mac Studio (Apple M5 Max)4bit—433.7 GB est.No
MacBook Pro (Apple M5 Max)4bit—433.7 GB est.No
Apple M2 Max4bit—433.7 GB est.No
Apple M1 Max4bit—433.7 GB est.No
Apple M4 Pro4bit—433.7 GB est.No
Mac mini (Apple M5 Pro)4bit—433.7 GB est.No
MacBook Pro (Apple M5 Pro)4bit—433.7 GB est.No
Apple M3 Pro4bit—433.7 GB est.No
Apple M1 Pro4bit—433.7 GB est.No
Apple M2 Pro4bit—433.7 GB est.No
Apple M44bit—433.7 GB est.No
iMac (Apple M4)4bit—433.7 GB est.No
Mac mini (Apple M6)4bit—433.7 GB est.No
MacBook Air (Apple M5)4bit—433.7 GB est.No
MacBook Pro (Apple M5)4bit—433.7 GB est.No
Apple M24bit—433.7 GB est.No
Apple M34bit—433.7 GB est.No
Apple M14bit—433.7 GB est.No
NVIDIA GeForce RTX 30904bit24 GB433.7 GB est.No
NVIDIA GeForce RTX 40904bit24 GB433.7 GB est.No
NVIDIA GeForce RTX 50904bit32 GB433.7 GB est.No
NVIDIA A100 80GB4bit80 GB433.7 GB est.No
NVIDIA H100 SXM4bit80 GB433.7 GB est.No
NVIDIA H100 NVL4bit94 GB433.7 GB est.No
NVIDIA DGX Spark4bit128 GB433.7 GB est.No
NVIDIA H2004bit141 GB433.7 GB est.No
NVIDIA H200 NVL4bit141 GB433.7 GB est.No
NVIDIA B2004bit180 GB433.7 GB est.No
AMD Instinct MI300X4bit192 GB433.7 GB est.No
AMD Instinct MI325X4bit256 GB433.7 GB est.No
Assumptions (7)
  • Estimated, not measured: weights = parameters × bytes/param × 1.15 runtime overhead (or the observed artifact file size when one is recorded).
  • bytes/param: 4bit = 0.5, 8bit = 1.0, fp16 = 2.0 (uniform quantization, no per-layer exceptions).
  • KV cache: 2 × layers × kv_heads × head_dim × 2 bytes × context × batch when the architecture is known; otherwise 0.5 GB per 8 192 tokens (× batch), independent of architecture (GQA/MLA models need less).
  • A model 'fits' when the estimate is at most the device memory minus 2 GB reserved for the OS and framework.
  • Mixture-of-experts models are estimated on total parameters (all experts must be resident); active parameters are ignored.
  • Device memory uses the largest configuration when several are listed (e.g. Apple silicon tiers).
  • Multi-GPU: device memories are summed; interconnect bandwidth, tensor-parallel replication and pipeline bubbles are not modelled.

Lineage

Open in Graph →
Explicit derived_from / fine_tuned_from / distilled_from relations stated by sources; artifacts collapsed by kind.
DESCENDANTS 0 · ARTIFACTS2 quantizationsartifacts · collapsed2 quantizations — artifacts · collapsedZ.ai GLM 5.2753.3B params · this modelZ.ai GLM 5.2 — 753.3B params · this model

    Versions & Artifacts2

    Version history

    Context window1 change

    11 Sept 2026→12 Sept 2026current

    Licensefirst observation only

    11 Sept 2026current

    Max outputfirst observation only

    11 Sept 2026current

    Opennessfirst observation only

    11 Sept 2026current

    Parametersfirst observation only

    11 Sept 2026current

    Each hop is a claim: click a value for its source, tier and observation time. Nothing is overwritten — a new observation closes the previous claim.

    Artifacts 2

    quantizations 2

    Papers2

    Timeline17

    Full timeline →

    Change history

    Temporal, append-only claims: a new observation closes the previous claim instead of overwriting it. Rewind the record with the as-of picker.
    2 claims · 1 propertiesShow all properties

    Aa median output tokens per secondmetric.aa_median_output_tokens_per_second2

    Claim history for Aa median output tokens per second
    ValueValid from → toStatusSourceConfidenceExtractor
    67.1 → currentcurrentArtificial AnalysisT2mediumdeterministic
    119.8 → currentcurrentArtificial AnalysisT2mediumdeterministic

    Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →

    Provenance

    Attributed facts

    44

    Source tiers

    T1T25 / 39

    Freshest observation

    3 h ago

    Conflicts

    7 flagged

    Source documents 10

    Source documents
    SourceDocumentTypeTierLast observedSnapshots
    Mistral AI docsdocs.mistral.ai/models model_docsT1· Official2 h ago1
    Mistral AI docsdocs.mistral.ai/models/zai-glm-5-2 model_pageT1· Official14 h ago1
    OpenRouter public model & pricing listingopenrouter.ai/api/v1/models catalogueT2· Quality secondary4 min ago11
    Artificial Analysisartificialanalysis.ai/leaderboards/models leaderboardT2· Quality secondary3 h ago2
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/models?author=zai-org&p=0&sort=downloads listingT2· Quality secondary3 h ago7
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/zai-org/GLM-5.2/raw/main/README.md model_cardT2· Quality secondary6 h ago1
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/zai-org/GLM-5.2 model_pageT2· Quality secondary7 h ago4
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/nvidia/GLM-5.2-NVFP4 model_pageT2· Quality secondary7 h ago4
    Hugging Face Hub (public pages, model cards, papers)huggingface.co/amd/GLM-5.2-MXFP4 model_pageT2· Quality secondary8 h ago6
    LiveBenchlivebench.ai/table_2026_06_25.csv leaderboardT2· Quality secondary12 h ago1

    Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.

    Data quality (76/100) measures how well AI Atlas knows this entity — completeness, primary-source ratio, freshness, conflicts — never how good the model is. Methodology →