Skip to main content
BenchLM
Advertise

Model comparison

Qwen3.7 Max vs Qwen3.8 Max

Updated August 3, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

Qwen3.7 Max

Alibaba

71.8/100

Supported · Public rank #13

90% interval 65.3–78.2

Qwen3.8 Max

Alibaba

65.4/100

Estimated · Public rank #31

90% interval 55.5–75.3

Qwen3.7 Max has the higher public score estimate, 71.77 versus 65.4, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.

8 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

    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

    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: 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
8
Qwen3.7 Max only
27
Qwen3.8 Max only
44
Like-for-like categories
0 / 8

4 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.

Coding

Directional only
Qwen3.7 Max
77.9
Qwen3.8 Max
67.7
Weighted basis
4 vs 1 rows
Reading
Directional only

Reasoning

Directional only
Qwen3.7 Max
90.4
Qwen3.8 Max
78.3
Weighted basis
1 vs 2 rows
Reading
Directional only

Knowledge

Directional only
Qwen3.7 Max
64.2
Qwen3.8 Max
50.2
Weighted basis
4 vs 2 rows
Reading
Directional only

Instruction following

Directional only
Qwen3.7 Max
84.4
Qwen3.8 Max
82.8
Weighted basis
2 vs 1 rows
Reading
Directional only

Agentic

Not comparable
Qwen3.7 Max
69.7
Qwen3.8 Max
86.1
Weighted basis
1 vs 1 rows
Reading
Not comparable

Math

Not comparable
Qwen3.7 Max
97.1
Qwen3.8 Max
Not measured
Weighted basis
1 vs 0 rows
Reading
Not comparable

Multilingual

Not comparable
Qwen3.7 Max
87.0
Qwen3.8 Max
Not measured
Weighted basis
1 vs 0 rows
Reading
Not comparable

Multimodal

Not comparable
Qwen3.7 Max
Not measured
Qwen3.8 Max
86.3
Weighted basis
0 vs 2 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

Qwen3.7 Max
API rate not published
Fits in one request
Qwen3.8 Max
API rate not published
Fits in one request

Qwen3.7 Max has no comparable published API token rate. Qwen3.8 Max has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

Qwen3.7 Max
API rate not published
Fits in one request
Qwen3.8 Max
API rate not published
Fits in one request

Qwen3.7 Max has no comparable published API token rate. Qwen3.8 Max has no comparable published API token rate.

Cache-heavy agent loop

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

Qwen3.7 Max
API rate not published
Fits in one request
Cached-input rate unavailable
Qwen3.8 Max
API rate not published
Fits in one request
Cached-input rate unavailable

Qwen3.7 Max has no comparable published API token rate. Qwen3.8 Max 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.

Cached-input rate

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

Qwen3.7 Max

No comparable hosted API rate

Qwen3.8 Max

No comparable hosted API rate

Alibaba Cloud Model Studio pricing

Documented inputs

Qwen3.7 Max

Not sourced

Qwen3.8 Max

Not sourced

Documented outputs

Qwen3.7 Max

Not sourced

Qwen3.8 Max

Not sourced

Provider availability

Qwen3.7 Max

Not sourced

Qwen3.8 Max

Not sourced

Reasoning profile

Qwen3.7 Max

Reasoning

Qwen3.8 Max

Reasoning

Weight access

Qwen3.7 Max

Proprietary

Qwen3.8 Max

Proprietary

License

Qwen3.7 Max

Proprietary

Qwen3.8 Max

Proprietary

Release date

Qwen3.7 Max

2026-05-16

Qwen3.8 Max

2026-08-03

If you already use one of these models
Deployment change
Both entries list Alibaba as the provider. Confirm endpoint, model ID, limits, and feature support before switching.
Quality signal
Qwen3.7 Max has the higher public score estimate, 71.77 versus 65.4, but the 90% score intervals overlap.
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 evidence79 rows

Agentic

  • Terminal-Bench 2.0

    Qwen3.7 Max69.7%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • QwenClawBench

    Qwen3.7 Max64.3%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • Claw-Eval

    Qwen3.7 Max65.2%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • BFCL v4

    Qwen3.7 Max75.0%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • MCP Atlas

    Qwen3.7 Max76.4%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • VITA-Bench

    Qwen3.7 Max47.9%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • HLE w/ tools

    Qwen3.7 Max53.5%
    Source
    Qwen3.8 Max56.2%
    Source

    Qwen3.8 Max leads this result

  • Gert Labs

    Qwen3.7 Max64.27%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • ResearchClawBench

    Qwen3.7 Max18.7%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • Terminal-Bench 2.1

    Qwen3.7 Max—
    Qwen3.8 Max86.6%
    Source

    Not directly comparable

  • CoWorkBench

    Qwen3.7 Max—
    Qwen3.8 Max74.8%
    Source

    Not directly comparable

  • JobBench

    Qwen3.7 Max—
    Qwen3.8 Max53.4%
    Source

    Not directly comparable

  • skillsBench

    Qwen3.7 Max—
    Qwen3.8 Max70.2%
    Source

    Not directly comparable

  • Agents' Last Exam

    Qwen3.7 Max—
    Qwen3.8 Max52.4%
    Source

    Not directly comparable

  • AutomationBench

    Qwen3.7 Max—
    Qwen3.8 Max27.3%
    Source

    Not directly comparable

  • Toolathlon-Verified

    Qwen3.7 Max—
    Qwen3.8 Max72.5%
    Source

    Not directly comparable

  • WideResearch

    Qwen3.7 Max—
    Qwen3.8 Max81.9%
    Source

    Not directly comparable

  • OSWorld-Verified

    Qwen3.7 Max—
    Qwen3.8 Max86.1%
    Source

    Not directly comparable

  • OSWorld 2.0

    Qwen3.7 Max—
    Qwen3.8 Max19.4%
    Source

    Not directly comparable

  • WebArena-Verified

    Qwen3.7 Max—
    Qwen3.8 Max66.8%
    Source

    Not directly comparable

  • AndroidWorld

    Qwen3.7 Max—
    Qwen3.8 Max85.3%
    Source

    Not directly comparable

  • MobileWorld

    Qwen3.7 Max—
    Qwen3.8 Max77.8%
    Source

    Not directly comparable

Coding

  • SWE-bench Verified

    Qwen3.7 Max80.4%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • SWE-bench Pro

    Qwen3.7 Max60.6%
    Source
    Qwen3.8 Max67.7%
    Source

    Qwen3.8 Max leads this result

  • SWE Multilingual

    Qwen3.7 Max78.3%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • NL2Repo

    Qwen3.7 Max47.2%
    Source
    Qwen3.8 Max55.9%
    Source

    Qwen3.8 Max leads this result

  • SciCode

    Qwen3.7 Max53.5%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • LiveCodeBench

    Qwen3.7 Max91.6%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • Terminal-Bench 2.0

    Qwen3.7 Max69.7%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • Terminal-Bench 2.1

    Qwen3.7 Max—
    Qwen3.8 Max86.6%
    Source

    Not directly comparable

  • deepSwe

    Qwen3.7 Max—
    Qwen3.8 Max56.6%
    Source

    Not directly comparable

  • FrontierSWE

    Qwen3.7 Max—
    Qwen3.8 Max73.5%
    Source

    Not directly comparable

  • MLS-Bench Lite

    Qwen3.7 Max—
    Qwen3.8 Max41.0%
    Source

    Not directly comparable

  • PaperBench

    Qwen3.7 Max—
    Qwen3.8 Max93.0%
    Source

    Not directly comparable

Reasoning

  • MRCRv2

    Qwen3.7 Max90.4%
    Source
    Qwen3.8 Max92.9%
    Source

    Qwen3.8 Max leads this result

  • CritPt

    Qwen3.7 Max13.4%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • LongBench v2

    Qwen3.7 Max—
    Qwen3.8 Max66.3%
    Source

    Not directly comparable

Knowledge

  • GPQA

    Qwen3.7 Max92.4%
    Source
    Qwen3.8 Max92.6%
    Source

    Qwen3.8 Max leads this result

  • GPQA-D

    Qwen3.7 Max92.4%
    Source
    Qwen3.8 Max92.6%
    Source

    Qwen3.8 Max leads this result

  • HLE

    Qwen3.7 Max41.4%
    Source
    Qwen3.8 Max43.6%
    Source

    Qwen3.8 Max leads this result

  • MMLU-Pro

    Qwen3.7 Max89.6%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • MMLU-Redux

    Qwen3.7 Max95%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • SuperGPQA

    Qwen3.7 Max73.6%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • MMMLU

    Qwen3.7 Max90.3%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • HLE w/o tools

    Qwen3.7 Max—
    Qwen3.8 Max43.6%
    Source

    Not directly comparable

Math

  • HMMT Feb 2026

    Qwen3.7 Max97.1%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • IMOAnswerBench

    Qwen3.7 Max90.0%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • Apex

    Qwen3.7 Max44.5%
    Source
    Qwen3.8 Max—

    Not directly comparable

Multilingual

  • MMLU-ProX

    Qwen3.7 Max87%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • NOVA-63

    Qwen3.7 Max59.0%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • INCLUDE

    Qwen3.7 Max86.2%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • MAXIFE

    Qwen3.7 Max89.2%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • PolyMath

    Qwen3.7 Max86.5%
    Source
    Qwen3.8 Max—

    Not directly comparable

Multimodal

  • MMMU-Pro

    Qwen3.7 Max—
    Qwen3.8 Max82.3%
    Source

    Not directly comparable

  • MathVision

    Qwen3.7 Max—
    Qwen3.8 Max95.2%
    Source

    Not directly comparable

  • MathVision w/ Python

    Qwen3.7 Max—
    Qwen3.8 Max97.7%
    Source

    Not directly comparable

  • BabyVision

    Qwen3.7 Max—
    Qwen3.8 Max82.0%
    Source

    Not directly comparable

  • BabyVision w/ Python

    Qwen3.7 Max—
    Qwen3.8 Max91.3%
    Source

    Not directly comparable

  • ZeroBench

    Qwen3.7 Max—
    Qwen3.8 Max24.0%
    Source

    Not directly comparable

  • ZeroBench w/ Python

    Qwen3.7 Max—
    Qwen3.8 Max49.0%
    Source

    Not directly comparable

  • MedXpertQA (MM)

    Qwen3.7 Max—
    Qwen3.8 Max80.4%
    Source

    Not directly comparable

  • ScreenSpot Pro

    Qwen3.7 Max—
    Qwen3.8 Max84.5%
    Source

    Not directly comparable

  • Vision2Web

    Qwen3.7 Max—
    Qwen3.8 Max69.0%
    Source

    Not directly comparable

  • CharXiv w/o tools

    Qwen3.7 Max—
    Qwen3.8 Max88.4%
    Source

    Not directly comparable

  • CharXiv

    Qwen3.7 Max—
    Qwen3.8 Max93.5%
    Source

    Not directly comparable

  • OmniDocBench 1.5

    Qwen3.7 Max—
    Qwen3.8 Max92.1%
    Source

    Not directly comparable

  • OCRBench V2

    Qwen3.7 Max—
    Qwen3.8 Max74.2%
    Source

    Not directly comparable

  • CC-OCR

    Qwen3.7 Max—
    Qwen3.8 Max79.6%
    Source

    Not directly comparable

  • RealWorldQA

    Qwen3.7 Max—
    Qwen3.8 Max88.0%
    Source

    Not directly comparable

  • ERQA

    Qwen3.7 Max—
    Qwen3.8 Max77.8%
    Source

    Not directly comparable

  • SimpleVQA

    Qwen3.7 Max—
    Qwen3.8 Max75.0%
    Source

    Not directly comparable

  • PerceptionBench

    Qwen3.7 Max—
    Qwen3.8 Max63.5%
    Source

    Not directly comparable

  • Video-MME (with subtitle)

    Qwen3.7 Max—
    Qwen3.8 Max90.4%
    Source

    Not directly comparable

  • VideoMMMU

    Qwen3.7 Max—
    Qwen3.8 Max88.7%
    Source

    Not directly comparable

  • MMVU

    Qwen3.7 Max—
    Qwen3.8 Max82.4%
    Source

    Not directly comparable

  • MLVU (M-Avg)

    Qwen3.7 Max—
    Qwen3.8 Max90.8%
    Source

    Not directly comparable

  • LVBench

    Qwen3.7 Max—
    Qwen3.8 Max81.8%
    Source

    Not directly comparable

Instruction following

  • IFEval

    Qwen3.7 Max94.3%
    Source
    Qwen3.8 Max—

    Not directly comparable

  • IFBench

    Qwen3.7 Max79.1%
    Source
    Qwen3.8 Max82.8%
    Source

    Qwen3.8 Max leads this result

Frequently asked questions

Which is better, Qwen3.7 Max or Qwen3.8 Max?

Qwen3.7 Max has the higher public score estimate, 71.77 versus 65.4, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.

Which is better for coding, Qwen3.7 Max or Qwen3.8 Max?

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, Qwen3.7 Max or Qwen3.8 Max?

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, Qwen3.7 Max or Qwen3.8 Max?

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, Qwen3.7 Max or Qwen3.8 Max?

Both models list the same context window, 1M.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated August 3, 2026

Watch Qwen3.7 Max vs Qwen3.8 Max

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

Read a sample issue

Join 2,000+ readers.