Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d263Policy30d65, 14 pendingSources1,013

GLM-5.2

Z.aiReleased Jun 16, 2026Current

Score basis

GLM-5.2 is capable in long context, coding, reasoning, and agentic tasks; and behind the leaders in factuality, instruction following, and math. Too few results yet to rate safety, multimodal tasks, or multilingual tasks.

Capability profile

Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.

Capable4 capabilities
  1. Long Context−14.0%1 of 3
  2. Coding−16.2%7 of 10
  3. Reasoning−18.0%6 of 6
  4. Agentic−24.2%4 of 7
Limited3 capabilities
  1. Factuality−25.6%3 of 4
  2. Instruction Following−34.0%2 of 3
  3. Math−43.7%5 of 5
Not rated3 capabilities

Too few results yet to rate Safety, Multimodal or Multilingual.

Price

$1.40input$4.40outputper million tokens

From Artificial Analysis · 8 providers tracked · All prices

Costs more than 73% of 331 priced models · 3:1 input-to-output blend, log scale

Evidence

123results on73benchmarks

  • 21 independently verified
  • 34 aggregator
  • 31 vendor-reported
  • 37 cross-referenced

From 23 sources · latest Oct 8, 2026 · How verification works

API features

Tool callingStructured outputsReasoning

As listed by OpenRouter

Research

35 papers reference GLM-5.2

GLM-5.2 benchmark results

123 results on 73 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.

Long Context

CapableFull long context ranking

14.0% behind the leader1 of 3 ranked benchmarks measured

Show 1 more long context resultHide 1 long context result

Coding

CapableFull coding ranking

16.2% behind the leader7 of 10 ranked benchmarks measured

Show 7 more coding resultsHide 7 coding results

Reasoning

CapableFull reasoning ranking

18.0% behind the leader6 of 6 ranked benchmarks measured

Show 12 more reasoning resultsHide 12 reasoning results

Agentic

CapableFull agentic ranking

24.2% behind the leader4 of 7 ranked benchmarks measured

Show 8 more agentic resultsHide 8 agentic results

Factuality

LimitedFull factuality ranking

25.6% behind the leader3 of 4 ranked benchmarks measured

Show 3 more factuality resultsHide 3 factuality results

Instruction Following

LimitedFull instruction following ranking

34.0% behind the leader2 of 3 ranked benchmarks measured

Show 1 more instruction following resultHide 1 instruction following result

Math

LimitedFull math ranking

43.7% behind the leader5 of 5 ranked benchmarks measured

Show 6 more math resultsHide 6 math results

More results

Benchmarks outside the capability baskets. They are not ranked against a leader.

Show 53 more resultsHide 53 results

GLM-5.2: common questions

Who makes GLM-5.2?

GLM-5.2 is made by Z.ai.

When was GLM-5.2 released?

GLM-5.2 was released on Jun 16, 2026, according to Artificial Analysis.

What is GLM-5.2 good at?

GLM-5.2 is capable in long context, coding, reasoning, and agentic tasks; and behind the leaders in factuality, instruction following, and math. Too few results yet to rate safety, multimodal tasks, or multilingual tasks.

How much does GLM-5.2 cost?

GLM-5.2 costs $1.40 per million input tokens and $4.40 per million output tokens, according to Artificial Analysis. We track its price at 8 providers. At a mix of three input tokens to one output token, it costs more than 73% of the 331 priced models we track.

How many benchmarks has GLM-5.2 been tested on?

We track 123 results for GLM-5.2 on 73 benchmarks from 23 sources, 21 of them independently verified. The latest was recorded on Oct 8, 2026.

Which API features does GLM-5.2 support?

OpenRouter lists tool calling, structured outputs, and reasoning for GLM-5.2.

About this record

Where GLM-5.2's numbers come from, and every name it appears under.

Tracked since
Jun 15, 2026
Newest source mention
Aug 28, 2026

Where the results come from

Verification: 123 scores · 21 independently verified · 34 aggregator-attributed · 37 vendor cross-reference · 31 vendor-reported. How these tiers are assigned

From 23 sources on 12 sites. Hugging Face supplies 49 of them; the 21 independently verified results come from 9 sites. Bars are coloured by trust tier.

  • huggingface.co49
  • artificialanalysis.ai36
  • api.llm-stats.com16
  • livebench.ai7
  • epoch.ai3
  • thinkingmachines.ai3
  • arcprize.org2
  • labs.scale.com2
  • matharena.ai2
  • datasets-server.huggingface.co1
  • lmarena.ai1
  • simple-bench.com1

Also known as

glm 5.2 maxglm 52glm-5-2glm-5-2-non-reasoningglm-5.2 (max)glm-5.2 (non-reasoning)glm-5.2 full open sourceglm-5.2:freeglm-5p2glm5.2glm52z.ai glm 5.2zai-org/glm-5.2zai-org/glm-5.2-fp8zhipu glm-5.2