GLM-5

Zhipu AI
Model overview

GLM-5

GLM-5 series model.

general LLMreasoningcodingchatcode generationenterprise access
Representative version

GLM-5

CompanyZhipu AI
ReleaseNot disclosed
Release typeModel release
Parameters753.9B
ArchitectureNot disclosed
ContextNot disclosed
Input modalitiestext / code
Output modalitiestext / code
Tool use / function callingNot disclosed
Structured output / reasoningConfirmed: Reasoning
Vision / audio / videoNot disclosed
Open weights / LicenseWeight status not disclosed / license not disclosed
zhipu-glm-5-phase-3-5 · License source pending verification
Vendor APIAvailable
zai-org/GLM-5 Hugging Face model cardzhipu-glm-5-phase-3-5 · 2026-05-25
OpenRouterOpenRouter listed
PricingOpenRouter · 2026-07-15: $0.95 / $3.15
input / output, per 1M tokens
BenchmarkLMArena 4 rows
Public signals2 runtime / access records / 1 community discussion / 1 activity snapshot
Use casesassistant chat / code generation / enterprise API / general text tasks
Claim evidenceRelease eventzai-org/GLM-5 Hugging Face model cardGLM-5
Model specs / weightsZ.AI GLM-5 documentationGLM-5 · GLM-5
OpenRouter

Pricing / context / availability

OpenRouterOpenRouter listed
OpenRouter listingz-ai/glm-5
Mapped versionGLM-5
zhipu-glm-5-phase-3-5
PricingOpenRouter · 2026-07-15: $0.95 / $3.15
input / output, per 1M tokens
Context202752
Speed/LatencyNo reliable public source
Snapshot time2026-07-15T20:11:41.551Z
SourceView source
Benchmark

Ranking evidence

SourceTaskRankScoreSampleTimeLink
LMArenaWebDev241,435.566,5742026-05-07View source
LMArenaText overall331,456.6320,2922026-05-07View source
LMArenaReasoning / hard prompts371,476.9811,6762026-05-07View source
LMArenaCoding471,493.794,7742026-05-07View source
Public activity counters

Cumulative Hugging Face model-card downloads / likes / spaces

PlatformEntryDownloadsLikesSpacesTimeSource
huggingfacezai-org/GLM-596,5972,1161002026-07-15T20:11:41.551ZView source
Operations

Runtime / access signals

TimeSignalOriginal textSource
06/16/2026, 03:00 AM UTC[Rate Limit Reached] Pro Yearly users!Screenshot2026-06-16at11-00-02 Hello, I'm a pro subscribe user and I found the rate limit quota too little. Anybody can solve this?Source
04/15/2026, 09:56 PM UTCLoading GLM-5 on 2 x H200 nodes with 8 GPUsI need to load GLM-5 on 2 x H200 nodes with 8 GPUs each. While I set tensor-parallel = 8 but when I set 'pipelineparallelsize' 2, I am receivi...Source
Evidence gaps

Public display boundary

Release date not disclosedNo linkable real reviewSpeed/latency not disclosed