Skip to main content
BenchLM
Advertise

Model comparison

Qwen3.8 Max vs Soofi S 30B-A3B

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

Qwen3.8 Max

Alibaba

65.4/100

Estimated · Public rank #31

90% interval 55.5–75.3

Soofi S 30B-A3B

Soofi Project

—

Evidence status unavailable

90% interval unavailable

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: 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
2
Qwen3.8 Max only
50
Soofi S 30B-A3B only
6
Like-for-like categories
0 / 8

1 category uses 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.

Knowledge

Directional only
Qwen3.8 Max
50.2
Soofi S 30B-A3B
49.9
Weighted basis
2 vs 2 rows
Reading
Directional only

Agentic

Not comparable
Qwen3.8 Max
86.1
Soofi S 30B-A3B
Not measured
Weighted basis
1 vs 0 rows
Reading
Not comparable

Coding

Not comparable
Qwen3.8 Max
67.7
Soofi S 30B-A3B
Not measured
Weighted basis
1 vs 0 rows
Reading
Not comparable

Reasoning

Not comparable
Qwen3.8 Max
78.3
Soofi S 30B-A3B
Not measured
Weighted basis
2 vs 0 rows
Reading
Not comparable

Math

Not comparable
Qwen3.8 Max
Not measured
Soofi S 30B-A3B
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Multilingual

Not comparable
Qwen3.8 Max
Not measured
Soofi S 30B-A3B
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Multimodal

Not comparable
Qwen3.8 Max
86.3
Soofi S 30B-A3B
Not measured
Weighted basis
2 vs 0 rows
Reading
Not comparable

Instruction following

Not comparable
Qwen3.8 Max
82.8
Soofi S 30B-A3B
Not measured
Weighted basis
1 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

Qwen3.8 Max
API rate not published
Fits in one request
Soofi S 30B-A3B
Self-hosted; infrastructure cost varies
Fits in one request

Qwen3.8 Max has no comparable published API token rate. Soofi S 30B-A3B has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

Qwen3.8 Max
API rate not published
Fits in one request
Soofi S 30B-A3B
Self-hosted; infrastructure cost varies
Fits in one request

Qwen3.8 Max has no comparable published API token rate. Soofi S 30B-A3B has no comparable published API token rate.

Cache-heavy agent loop

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

Qwen3.8 Max
API rate not published
Fits in one request
Cached-input rate unavailable
Soofi S 30B-A3B
Self-hosted; infrastructure cost varies
Fits in one request
Cached-input rate unavailable

Qwen3.8 Max has no comparable published API token rate. Soofi S 30B-A3B 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.

Qwen3.8 Max

No comparable hosted API rate

Alibaba Cloud Model Studio pricing

Soofi S 30B-A3B

No comparable hosted API rate

Documented inputs

Qwen3.8 Max

Not sourced

Soofi S 30B-A3B

Not sourced

Documented outputs

Qwen3.8 Max

Not sourced

Soofi S 30B-A3B

Not sourced

Provider availability

Qwen3.8 Max

Not sourced

Soofi S 30B-A3B

Not sourced

Reasoning profile

Qwen3.8 Max

Reasoning

Soofi S 30B-A3B

Non-Reasoning

Weight access

Qwen3.8 Max

Proprietary

Soofi S 30B-A3B

Open Weight

License

Qwen3.8 Max

Proprietary

Soofi S 30B-A3B

Open Weight

Release date

Qwen3.8 Max

2026-08-03

Soofi S 30B-A3B

2026-07-10

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
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 evidence58 rows

Agentic

  • Terminal-Bench 2.1

    Qwen3.8 Max86.6%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • CoWorkBench

    Qwen3.8 Max74.8%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • JobBench

    Qwen3.8 Max53.4%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • skillsBench

    Qwen3.8 Max70.2%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • Agents' Last Exam

    Qwen3.8 Max52.4%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • AutomationBench

    Qwen3.8 Max27.3%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • Toolathlon-Verified

    Qwen3.8 Max72.5%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • WideResearch

    Qwen3.8 Max81.9%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • HLE w/ tools

    Qwen3.8 Max56.2%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • OSWorld-Verified

    Qwen3.8 Max86.1%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • OSWorld 2.0

    Qwen3.8 Max19.4%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • WebArena-Verified

    Qwen3.8 Max66.8%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • AndroidWorld

    Qwen3.8 Max85.3%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MobileWorld

    Qwen3.8 Max77.8%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

Coding

  • Terminal-Bench 2.1

    Qwen3.8 Max86.6%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • SWE-bench Pro

    Qwen3.8 Max67.7%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • deepSwe

    Qwen3.8 Max56.6%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • NL2Repo

    Qwen3.8 Max55.9%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • FrontierSWE

    Qwen3.8 Max73.5%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MLS-Bench Lite

    Qwen3.8 Max41.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • PaperBench

    Qwen3.8 Max93.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • HumanEval

    Qwen3.8 Max—
    Soofi S 30B-A3B73.8%
    Source

    Not directly comparable

Reasoning

  • MRCRv2

    Qwen3.8 Max92.9%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • LongBench v2

    Qwen3.8 Max66.3%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • BBH

    Qwen3.8 Max—
    Soofi S 30B-A3B78.8%
    Source

    Not directly comparable

  • DROP

    Qwen3.8 Max—
    Soofi S 30B-A3B66.5%
    Source

    Not directly comparable

Knowledge

  • GPQA

    Qwen3.8 Max92.6%
    Source
    Soofi S 30B-A3B43.4%
    Source

    Qwen3.8 Max leads this result

  • GPQA-D

    Qwen3.8 Max92.6%
    Source
    Soofi S 30B-A3B43.4%
    Source

    Qwen3.8 Max leads this result

  • HLE

    Qwen3.8 Max43.6%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • HLE w/o tools

    Qwen3.8 Max43.6%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MMLU-Pro

    Qwen3.8 Max—
    Soofi S 30B-A3B51.4%
    Source

    Not directly comparable

  • AGIEval

    Qwen3.8 Max—
    Soofi S 30B-A3B66.9%
    Source

    Not directly comparable

Math

  • GSM8K

    Qwen3.8 Max—
    Soofi S 30B-A3B86.1%
    Source

    Not directly comparable

Multimodal

  • MMMU-Pro

    Qwen3.8 Max82.3%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MathVision

    Qwen3.8 Max95.2%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MathVision w/ Python

    Qwen3.8 Max97.7%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • BabyVision

    Qwen3.8 Max82.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • BabyVision w/ Python

    Qwen3.8 Max91.3%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • ZeroBench

    Qwen3.8 Max24.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • ZeroBench w/ Python

    Qwen3.8 Max49.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MedXpertQA (MM)

    Qwen3.8 Max80.4%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • ScreenSpot Pro

    Qwen3.8 Max84.5%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • Vision2Web

    Qwen3.8 Max69.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • CharXiv w/o tools

    Qwen3.8 Max88.4%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • CharXiv

    Qwen3.8 Max93.5%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • OmniDocBench 1.5

    Qwen3.8 Max92.1%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • OCRBench V2

    Qwen3.8 Max74.2%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • CC-OCR

    Qwen3.8 Max79.6%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • RealWorldQA

    Qwen3.8 Max88.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • ERQA

    Qwen3.8 Max77.8%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • SimpleVQA

    Qwen3.8 Max75.0%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • PerceptionBench

    Qwen3.8 Max63.5%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • Video-MME (with subtitle)

    Qwen3.8 Max90.4%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • VideoMMMU

    Qwen3.8 Max88.7%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MMVU

    Qwen3.8 Max82.4%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • MLVU (M-Avg)

    Qwen3.8 Max90.8%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

  • LVBench

    Qwen3.8 Max81.8%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

Instruction following

  • IFBench

    Qwen3.8 Max82.8%
    Source
    Soofi S 30B-A3B—

    Not directly comparable

Frequently asked questions

Which is better, Qwen3.8 Max or Soofi S 30B-A3B?

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, Qwen3.8 Max or Soofi S 30B-A3B?

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, Qwen3.8 Max or Soofi S 30B-A3B?

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, Qwen3.8 Max or Soofi S 30B-A3B?

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, Qwen3.8 Max or Soofi S 30B-A3B?

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 Qwen3.8 Max vs Soofi S 30B-A3B

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

Read a sample issue

Join 2,000+ readers.