Skip to main content
BenchLM
Advertise

Model comparison

GPT-5.4 vs Qwen 3.6 Max (preview)

Updated July 29, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

GPT-5.4

OpenAI

73.3/100

Supported · Public rank #10

90% interval 70.3–76.2

Qwen 3.6 Max (preview)

Alibaba

59.3/100

Supported · Public rank #56

90% interval 40.4–78.2

GPT-5.4 has the higher public score estimate, 73.25 versus 59.29, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.

4 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.

  • Long documents

    Prompts that approach the documented context limit

    GPT-5.4

    GPT-5.4 has the larger documented context window.

    Confidence: documented

Show secondary and unsupported calls

  • Coding work

    Code generation, repair, and software-engineering tasks

    Not enough matched evidence

    The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.

    Confidence: limited

  • Agentic work

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

    Not enough matched evidence

    The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.

    Confidence: limited

  • 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: listed-rates

  • 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
4
GPT-5.4 only
33
Qwen 3.6 Max (preview) only
5
Like-for-like categories
1 / 8

2 categories use different evidence sets. Those rows remain visible for coverage context but do not name a winner.

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.

Math

Like-for-like
GPT-5.4
42.5
Qwen 3.6 Max (preview)
18.4
Weighted basis
2 vs 2 rows
Reading
GPT-5.4 leads

Agentic

Directional only
GPT-5.4
77.2
Qwen 3.6 Max (preview)
65.4
Weighted basis
3 vs 1 rows
Reading
Directional only

Coding

Directional only
GPT-5.4
57.7
Qwen 3.6 Max (preview)
51.0
Weighted basis
1 vs 2 rows
Reading
Directional only

Reasoning

Not comparable
GPT-5.4
74.0
Qwen 3.6 Max (preview)
Not measured
Weighted basis
1 vs 0 rows
Reading
Not comparable

Knowledge

Not comparable
GPT-5.4
57.6
Qwen 3.6 Max (preview)
73.9
Weighted basis
2 vs 1 rows
Reading
Not comparable

Multilingual

Not comparable
GPT-5.4
Not measured
Qwen 3.6 Max (preview)
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Multimodal

Not comparable
GPT-5.4
73.2
Qwen 3.6 Max (preview)
Not measured
Weighted basis
3 vs 0 rows
Reading
Not comparable

Instruction following

Not comparable
GPT-5.4
Not measured
Qwen 3.6 Max (preview)
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.

Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.

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

GPT-5.4
$0.01
Fits in one request
Qwen 3.6 Max (preview)
API rate not published
Fits in one request

Qwen 3.6 Max (preview) has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

GPT-5.4
$0.17
Fits in one request
Qwen 3.6 Max (preview)
API rate not published
Fits in one request

Qwen 3.6 Max (preview) has no comparable published API token rate.

Cache-heavy agent loop

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

GPT-5.4
$0.25
Fits in one request
Qwen 3.6 Max (preview)
API rate not published
Fits in one request
Cached-input rate unavailable

Qwen 3.6 Max (preview) 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.

Context window

Maximum documented context; output-token limits may be lower.

GPT-5.4

1.05M

OpenAI pricing

Qwen 3.6 Max (preview)

256K

API model ID

GPT-5.4

gpt-5.4

OpenAI pricing

Qwen 3.6 Max (preview)

Not sourced

Cached-input rate

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

GPT-5.4

$0.25 per 1M cached input tokens

OpenAI pricing

Qwen 3.6 Max (preview)

No comparable hosted API rate

Documented inputs

GPT-5.4

Not sourced

Qwen 3.6 Max (preview)

Not sourced

Documented outputs

GPT-5.4

Not sourced

Qwen 3.6 Max (preview)

Not sourced

Provider availability

GPT-5.4

Not sourced

Qwen 3.6 Max (preview)

Not sourced

Reasoning profile

GPT-5.4

Reasoning

Qwen 3.6 Max (preview)

Reasoning

Weight access

GPT-5.4

Proprietary

Qwen 3.6 Max (preview)

Proprietary

License

GPT-5.4

Proprietary

Qwen 3.6 Max (preview)

Proprietary

Release date

GPT-5.4

2026-03-05

Qwen 3.6 Max (preview)

2026-04-20

If you already use one of these models
Deployment change
The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
Quality signal
GPT-5.4 has the higher public score estimate, 73.25 versus 59.29, but the 90% score intervals overlap.
Workload cost
A complete comparable API-rate estimate is not available for both models.
Context tradeoff
GPT-5.4 has the larger documented window (1.05M).

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 evidence42 rows

Agentic

  • Terminal-Bench 2.0

    GPT-5.475.1%
    Source
    Qwen 3.6 Max (preview)65.4%
    Source

    GPT-5.4 leads this result

  • CyberGym

    GPT-5.479.0%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • BrowseComp

    GPT-5.482.7%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • OSWorld-Verified

    GPT-5.475%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • MCP Atlas

    GPT-5.470.6%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • Toolathlon

    GPT-5.454.6%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • τ²-bench results

    GPT-5.498.9%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • Claw-Eval

    GPT-5.460.3%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • DeepSearchQA

    GPT-5.473.6%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • Gert Labs

    GPT-5.464.89%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • ResearchClawBench

    GPT-5.415.3%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • JobBench

    GPT-5.438.9%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • ExploitGym

    GPT-5.46.0%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • QwenClawBench

    GPT-5.4—
    Qwen 3.6 Max (preview)59.0%
    Source

    Not directly comparable

Coding

  • LiveCodeBench Pro

    GPT-5.487.5%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • SWE-bench Pro

    GPT-5.457.7%
    Source
    Qwen 3.6 Max (preview)57.3%
    Source

    GPT-5.4 leads this result

  • React Native Evals

    GPT-5.485.3%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • Vibe Code Bench

    GPT-5.467.42%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • SciCode

    GPT-5.4—
    Qwen 3.6 Max (preview)47%
    Source

    Not directly comparable

  • NL2Repo

    GPT-5.4—
    Qwen 3.6 Max (preview)42.9%
    Source

    Not directly comparable

  • Terminal-Bench 2.0

    GPT-5.4—
    Qwen 3.6 Max (preview)65.4%
    Source

    Not directly comparable

Reasoning

  • ARC-AGI-2

    GPT-5.474.0%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • ARC-AGI-3

    GPT-5.40.2%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

Knowledge

  • GPQA

    GPT-5.492.8%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • HLE

    GPT-5.452.1%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • HLE w/o tools

    GPT-5.439.8%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • GPQA-D

    GPT-5.492.8%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • HealthBench Hard

    GPT-5.440.1%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • MedXpertQA (Text)

    GPT-5.459.6%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • HealthBench Professional

    GPT-5.448.1%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • SuperGPQA

    GPT-5.4—
    Qwen 3.6 Max (preview)73.9%
    Source

    Not directly comparable

Math

  • FrontierMath v2 (Tiers 1-3)

    Shared source
    GPT-5.447.600%
    Qwen 3.6 Max (preview)23.103%

    GPT-5.4 leads this result

  • FrontierMath v2 (Tier 4)

    Shared source
    GPT-5.427.100%
    Qwen 3.6 Max (preview)4.167%

    GPT-5.4 leads this result

Multimodal

  • MMMU-Pro

    GPT-5.481.2%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • OfficeQA Pro

    GPT-5.453.2%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • MMMU-Pro w/ Python

    GPT-5.482.1%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • CharXiv

    GPT-5.482.8%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • ERQA

    GPT-5.465.4%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • SimpleVQA

    GPT-5.461.1%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • ScreenSpot Pro

    GPT-5.485.4%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • ZeroBench

    GPT-5.441.0%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

  • MedXpertQA (MM)

    GPT-5.477.1%
    Source
    Qwen 3.6 Max (preview)—

    Not directly comparable

Frequently asked questions

Which is better, GPT-5.4 or Qwen 3.6 Max (preview)?

GPT-5.4 has the higher public score estimate, 73.25 versus 59.29, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.

Which is better for coding, GPT-5.4 or Qwen 3.6 Max (preview)?

The current coding averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.

Which is better for agentic tasks, GPT-5.4 or Qwen 3.6 Max (preview)?

The current agentic tasks averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.

Which costs less, GPT-5.4 or Qwen 3.6 Max (preview)?

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, GPT-5.4 or Qwen 3.6 Max (preview)?

GPT-5.4 has the larger documented context window: 1.05M, compared with 256K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated July 29, 2026

Watch GPT-5.4 vs Qwen 3.6 Max (preview)

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

Read a sample issue

Join 2,000+ readers.