DataLearner logo
GL

GLM-5.2

Reasoning modelCoding modelGLMGLM-5

GLM-5.2

Release date: 2026-06-13Updated: 2026-08-22Views: 19,210
Live demoGitHubHugging FaceCompare
Parameters
753.33B
Context length
1M
Multilingual
Supported
Reasoning ability
5/5

GLM-5.2 is a reasoning model from Zhipu AI, released on 2026-06-13. It accepts text input and returns text output. The cataloged parameter count is 753.33B, with 40B active parameters per inference. The recorded context window is 1M, and the recorded maximum output is 128K. Cataloged capabilities include Reasoning model, Multilingual, and Coding model. The checkpoint is listed under the MIT license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

GLM-5.2

Model basics

Reasoning traces
Supported
Thinking modes
Thinking Level · High (Default)Standard ModeThinking Level · Max
Context length
1M tokens
Max output length
128K tokens
Model type
Reasoning model
Modality (in / out)
Text → Text
Release date
2026-06-13
Model file size
About 1.5 TB (BF16 safetensors)
MoE architecture
Yes
Total params / Active params
753.33B / 40B
Knowledge cutoff
No data
GLM-5.2

Open source & experience

Code license
MIT License
Weights license
MIT License- Commercial use permitted
GitHub repo
github.com/zai-org/GLM-5
Hugging Face
huggingface.co/zai-org/GLM-5.2
Live demo
chat.z.ai
GLM-5.2

Official resources

Paper
Paper
DataLearnerAI blog
N/A
GLM-5.2

API details

API speed
3/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Learn about pricing modes
Standard
TypeConditionInputOutput
Text-$1.40/ 1M tokens$4.40/ 1M tokens
Cache PricingPrompt Cache
TypeTTLWriteRead
Text-—$0.260/ 1M tokens

“—” means the modality is not billed in that direction, or the vendor has not published a price for it.

GLM-5.2

Benchmark Results

GLM-5.2 currently shows benchmark results led by τ²-Bench - Telecom (3 / 264, score 99.10), HLE (27 / 563, score 54.70), IMO-AnswerBench (2 / 24, score 91). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

7 evaluations
Benchmark / mode
Score
Rank/total
LiveBench
Standard Mode
73.18
21 / 117
HLE
Standard Mode
9.80
392 / 563
HLE
Thinking Mode
40.50
128 / 563
HLE
Thinking ModeTools
54.70
27 / 563
HLE
Max
41.10
123 / 563
CritPt
Standard Mode
3.10
110 / 200
CritPt
Max
20.90
33 / 200

General Evaluation

4 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Standard Mode
71.21
310 / 462
GPQA Diamond
Low
87.88
120 / 462
GPQA Diamond
Thinking Mode
91.20
63 / 462
GPQA Diamond
Max
91.86
57 / 462

Writing and Creative Capabilities

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
Standard Mode
1752.80
22 / 106

Common Sense Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
SimpleBench
Standard Mode
58.80
36 / 93

Coding and Software Engineer

12 evaluations
Benchmark / mode
Score
Rank/total
Text Arena (Coding)
Max
1593.25
5 / 35
FrontierSWE
MaxTools
74.40
3 / 4
WeirdML v2
HighTools
67.31
21 / 52
Program Bench
Thinking ModeTools
63.70
2 / 11
SWE-Bench Pro - Public
Thinking ModeTools
62.10
12 / 62
SciCode
Max
51.20
63 / 130
NL2Repo-Bench
Thinking ModeTools
48.90
13 / 16
DeepSWE
HighTools
36.28
74 / 85
DeepSWE
MaxTools
43.78
68 / 85
DeepSWE
DeepTools
44
67 / 85
PostTrain Bench
MaxTools
34.30
5 / 5
SWE-Marathon
MaxTools
13
6 / 6

Agent Level Benchmark

5 evaluations
Benchmark / mode
Score
Rank/total
τ²-Bench - Telecom
MaxTools
99.10
3 / 264
Terminal Bench Hard
MaxTools
50.80
22 / 244
τ³-Banking
Standard ModeTools
16.70
100 / 164
τ³-Banking
MaxTools
34.60
48 / 164
τ³-Banking
Extra-HighTools
37.11
40 / 164

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
Max
73.30
44 / 282

AI Agent - Tool Usage

6 evaluations
Benchmark / mode
Score
Rank/total
Terminal-Bench 2.1
Standard ModeTools
51.70
127 / 192
Terminal-Bench 2.1
HighTools
81
50 / 192
Terminal-Bench 2.1
MaxTools
77.90
68 / 192
MCP-Atlas
Thinking ModeTools
77.80
17 / 43
Tool Decathlon
Thinking ModeTools
48.20
4 / 10
Terminal-Bench 4.0
MaxTools
1
74 / 86

Text Embedding

1 evaluations
Benchmark / mode
Score
Rank/total
Context Arena
Max
72.34
58 / 126

Math and Reasoning

6 evaluations
Benchmark / mode
Score
Rank/total
AIME 2026
Thinking Mode
99.20
3 / 29
IMO-AnswerBench
Thinking Mode
91
2 / 24
FrontierMath v2
Standard Mode
42.46
34 / 58
FrontierMath v2
Low
54.74
28 / 58
FrontierMath v2
Max
59.21
22 / 58
FrontierMath Tier 4 v2
Max
29.27
20 / 42

Long Context

2 evaluations
Benchmark / mode
Score
Rank/total
AA-LCR
Standard Mode
42.30
154 / 170
AA-LCR
Max
78.30
63 / 170

Productivity Knowledge

4 evaluations
Benchmark / mode
Score
Rank/total
GDPval-AA v2
Standard ModeTools
1306
61 / 105
GDPval-AA v2
MaxTools
1406
46 / 105
AA-Briefcase
MaxTools
1231
42 / 83
Harvey Lab-AA
MaxTools
90.97
12 / 43

Multimodal Understanding

1 evaluations
Benchmark / mode
Score
Rank/total
GDP.pdf
Max
10.40
78 / 118

Claw-style Agent Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
PinchBench v2
unknown
86.98
5 / 45
View benchmark analysisCompare with other models

Compare with other models

GLM-5.2

Publisher

GLM-5.2

Model Overview

GLM-5.2 is a reasoning model from Zhipu AI, released on 2026-06-13.

It accepts text input and produces text output. Its cataloged capabilities include Reasoning model, Multilingual, and Coding model. The cataloged parameter count is 753.33B, with 40B active parameters per inference. The recorded context window is 1M, and the recorded maximum output is 128K.

The checkpoint is listed under the MIT license, with the license link included in the model references. The page records 3 API pricing rules from the listed provider; current provider pricing and conditions should be checked before deployment. The evaluation section contains 10 cataloged benchmark results with their recorded modes and scores. The page links 4 release, model-card, repository, or provider references for checking the underlying claims. Specifications, availability, and prices can change; undisclosed values are intentionally left unstated.

GLM-5.2

FAQ

What is GLM-5.2?

GLM-5.2 is a reasoning model from Zhipu AI, released on 2026-06-13. It accepts text input and returns text output. The cataloged parameter count is 753.33B, with 40B active parameters per inference. The recorded context window is 1M, and the recorded maximum output is 128K. Cataloged capabilities include Reasoning model, Multilingual, and Coding model. The checkpoint is listed under the MIT license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

What input and output modalities does GLM-5.2 support?

The current model record lists text as input and text as output.

What are the main recorded specifications for GLM-5.2?

The cataloged parameter count is 753.33B, with 40B active parameters per inference. The recorded context window is 1M, and the recorded maximum output is 128K. Fields without a source-backed value remain undisclosed.

Does GLM-5.2 have API pricing?

The page records 3 API pricing rules from the listed provider; current provider pricing and conditions should be checked before deployment.

Are benchmark results available for GLM-5.2?

The evaluation section contains 10 cataloged benchmark results with their recorded modes and scores. Compare only results that use the same benchmark version and evaluation mode.

Is GLM-5.2 open source?

The checkpoint is listed under the MIT license, with the license link included in the model references. Review the linked license text before commercial or derivative use.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code