Skip to main content
Radar

Five or fewer confirmed AI changes, with original sources, on mornings when something changed.A free source-linked morning brief.

Start free brief
BenchLM
Model A
GLM-5.2

Z.AI

63.2/100

Estimated · Public rank #44

90% interval 48.2–78.2

GLM-5.2 vs GLM-5.3

Updated August 14, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload. This is a same-family comparison, so migration details appear when the source data supports them.

Model B
GLM-5.3

Z.AI

—

Evidence status unavailable

90% interval unavailable

Decision reading

At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.

2 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.

Which one for your work

Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.

No workload recommendation clears the current evidence threshold.

Use the matched evidence, workload costs, and sourced specifications below instead of treating a point score as a universal answer.

Show secondary and unsupported calls

  • Coding work

    Code generation, repair, and software-engineering tasks

    Not enough matched evidence

    No shared weighted benchmark basis supports a winner.

    Confidence: limited

  • Agentic work

    Tool use, computer use, and multi-step task completion

    Not enough matched evidence

    No shared weighted benchmark basis supports a winner.

    Confidence: limited

  • Long documents

    Prompts that approach the documented context limit

    No clear pick

    The documented context windows are equal.

    Confidence: documented

  • Chat turn cost

    1K fresh input + 500 output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: listed-rates

  • Cache-heavy agent loop cost

    200K cached + 20K fresh input + 10K output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: rate-fallback

  • Repository review cost

    50K fresh input + 3K output tokens

    Not enough matched evidence

    A complete comparable API-rate estimate is not available for both models.

    Confidence: listed-rates

What is actually comparable

Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.

Shared results
2
GLM-5.2 only
18
GLM-5.3 only
14
Like-for-like categories
0 / 8

Category results, on a stated basis

Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.

Agentic

Not comparable
GLM-5.2
81.0
GLM-5.3
Not measured
Weighted basis
1 vs 0 rows
Reading
Not comparable

Coding

Not comparable
GLM-5.2
62.1
GLM-5.3
Not measured
Weighted basis
1 vs 0 rows
Reading
Not comparable

Reasoning

Not comparable
GLM-5.2
Not measured
GLM-5.3
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Knowledge

Not comparable
GLM-5.2
59.6
GLM-5.3
Not measured
Weighted basis
2 vs 0 rows
Reading
Not comparable

Math

Not comparable
GLM-5.2
95.9
GLM-5.3
Not measured
Weighted basis
2 vs 0 rows
Reading
Not comparable

Multilingual

Not comparable
GLM-5.2
Not measured
GLM-5.3
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Multimodal

Not comparable
GLM-5.2
Not measured
GLM-5.3
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Instruction following

Not comparable
GLM-5.2
Not measured
GLM-5.3
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Shape of the matched evidence

Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.

A shared-evidence shape is not available.

BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.

What each workload costs

Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.

Chat turn

1K fresh input + 500 output tokens

GLM-5.2
$0.0036
Fits in one request
GLM-5.3
API rate not published
Fits in one request

GLM-5.3 has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

GLM-5.2
$0.0832
Fits in one request
GLM-5.3
API rate not published
Fits in one request

GLM-5.3 has no comparable published API token rate.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

GLM-5.2
$0.352
Fits in one request
Cached input priced at the published list-input rate
GLM-5.3
API rate not published
Fits in one request
Cached-input rate unavailable

GLM-5.2 has no published cached-input rate, so cached tokens use its listed input rate. GLM-5.3 has no comparable published API token rate.

Specification differences

Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.

API model ID

GLM-5.2

Not sourced

GLM-5.3

Not sourced

Cached-input rate

A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.

GLM-5.2

Not published

GLM-5.3

No comparable hosted API rate

Z.AI GLM-5.3 launch post

Documented inputs

GLM-5.2

Not sourced

GLM-5.3

Not sourced

Documented outputs

GLM-5.2

Not sourced

GLM-5.3

Not sourced

Provider availability

GLM-5.2

Not sourced

GLM-5.3

Not sourced

Reasoning profile

GLM-5.2

Reasoning

GLM-5.3

Reasoning

Weight access

GLM-5.2

Open Weight

GLM-5.3

Proprietary

License

GLM-5.2

Open Weight

GLM-5.3

Proprietary

Release date

GLM-5.2

2026-06-16

GLM-5.3

2026-08-14

If you are considering the documented upgrade path
Deployment change
Both entries list Z.AI as the provider. Confirm endpoint, model ID, limits, and feature support before switching.
Quality signal
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
Workload cost
A complete comparable API-rate estimate is not available for both models.
Context tradeoff
Both models list 1M.

Run the same representative tasks against both endpoints before changing production traffic.

Benchmark evidence

The full public result ledger is available for audit without forcing a wide desktop table onto a phone.

Browse raw public benchmark evidence34 rows

Agentic

  • Terminal-Bench 3.0

    GLM-5.24.6%
    Source
    GLM-5.3—

    Not directly comparable

  • Terminal-Bench 2.0

    GLM-5.281%
    Source
    GLM-5.3—

    Not directly comparable

  • MCP Atlas

    GLM-5.276.8%
    Source
    GLM-5.3—

    Not directly comparable

  • Toolathlon

    GLM-5.248.2%
    Source
    GLM-5.3—

    Not directly comparable

  • ResearchClawBench

    GLM-5.220.7%
    Source
    GLM-5.3—

    Not directly comparable

  • Terminal-Bench 2.1

    GLM-5.2—
    GLM-5.388.2%
    Source

    Not directly comparable

  • terminalBench3

    GLM-5.2—
    GLM-5.328.3%
    Source

    Not directly comparable

  • CyberGym

    GLM-5.2—
    GLM-5.384.5%
    Source

    Not directly comparable

  • ExploitGym

    GLM-5.2—
    GLM-5.315.0%
    Source

    Not directly comparable

  • Toolathlon-Verified

    GLM-5.2—
    GLM-5.373.0%
    Source

    Not directly comparable

  • AutomationBench

    GLM-5.2—
    GLM-5.348.2%
    Source

    Not directly comparable

  • Agents' Last Exam

    GLM-5.2—
    GLM-5.328.5%
    Source

    Not directly comparable

  • HLE w/ tools

    GLM-5.2—
    GLM-5.362.5%
    Source

    Not directly comparable

Coding

  • SWE-bench Pro

    GLM-5.262.1%
    Source
    GLM-5.3—

    Not directly comparable

  • NL2Repo

    GLM-5.248.9%
    Source
    GLM-5.358%
    Source

    GLM-5.3 leads this result

  • Terminal-Bench 2.0

    GLM-5.281.0%
    Source
    GLM-5.3—

    Not directly comparable

  • ProgramBench

    GLM-5.263.7%
    Source
    GLM-5.319.0%
    Source

    GLM-5.2 leads this result

  • cursorBench32

    GLM-5.255.0%
    Source
    GLM-5.3—

    Not directly comparable

  • EEBench

    GLM-5.216.7%
    Source
    GLM-5.3—

    Not directly comparable

  • Terminal-Bench 2.1

    GLM-5.2—
    GLM-5.388.2%
    Source

    Not directly comparable

  • terminalBench3

    GLM-5.2—
    GLM-5.328.3%
    Source

    Not directly comparable

  • deepSwe

    GLM-5.2—
    GLM-5.366.9%
    Source

    Not directly comparable

  • FrontierSWE

    GLM-5.2—
    GLM-5.378.1%
    Source

    Not directly comparable

  • sweMarathon

    GLM-5.2—
    GLM-5.342.5%
    Source

    Not directly comparable

  • PostTrain Bench

    GLM-5.2—
    GLM-5.339.8%
    Source

    Not directly comparable

Reasoning

  • CritPt

    GLM-5.220.9%
    Source
    GLM-5.3—

    Not directly comparable

Knowledge

  • GPQA

    GLM-5.291.2%
    Source
    GLM-5.3—

    Not directly comparable

  • GPQA-D

    GLM-5.291.2%
    Source
    GLM-5.3—

    Not directly comparable

  • HLE

    GLM-5.254.7%
    Source
    GLM-5.3—

    Not directly comparable

  • HLE w/o tools

    GLM-5.240.5%
    Source
    GLM-5.3—

    Not directly comparable

Math

  • AIME26

    GLM-5.299.2%
    Source
    GLM-5.3—

    Not directly comparable

  • HMMT Nov 2025

    GLM-5.294.4%
    Source
    GLM-5.3—

    Not directly comparable

  • HMMT Feb 2026

    GLM-5.292.5%
    Source
    GLM-5.3—

    Not directly comparable

  • MMAnswerBench

    GLM-5.291.0%
    Source
    GLM-5.3—

    Not directly comparable

Frequently asked questions

Which is better, GLM-5.2 or GLM-5.3?

At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. The page therefore keeps the decision tied to the specific documented workload.

Which is better for coding, GLM-5.2 or GLM-5.3?

The published evidence does not provide a shared weighted coding basis for both models, so BenchLM does not name a coding winner.

Which is better for agentic tasks, GLM-5.2 or GLM-5.3?

The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.

Which costs less, GLM-5.2 or GLM-5.3?

Both models do not have comparable published API token rates, so this page does not name a universal price winner.

Which has the larger context window, GLM-5.2 or GLM-5.3?

Both models list the same context window, 1M.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated August 14, 2026

Watch GLM-5.2 vs GLM-5.3

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 2,000+ readers.