Skip to main content
BenchLM
Data

Qwen3.8 Max vs Beam 1.0 501B-A23B FP8

Updated October 16, 2026. We do not rank this pair: at least one has no public score. Public scores include evidence status and uncertainty.

Share or export
Share on XLinkedInAPI/MCP

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. 9 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.

Model A
Alibaba logo
Qwen3.8 Max

Alibaba

69.65/100

Supported · Public rank #19

90% interval 65.6–73.7

Model B
Beam 1.0 501B-A23B FP8

Reflection AI

—

Evidence status unavailable

90% interval unavailable

Shared results
9
Qwen3.8 Max only
57
Beam 1.0 501B-A23B FP8 only
10
Like-for-like categories
1 / 8
Supported: Qwen3.8 MaxHow the comparison works

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

    Beam 1.0 501B-A23B FP8 is not ranked on the public lane for coding, so no winner is named for coding.

    Confidence: limited

  • Agentic work

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

    Not enough matched evidence

    Beam 1.0 501B-A23B FP8 is not ranked on the public lane for agentic, so no winner is named for agentic.

    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

Which one for a specific job

Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.

The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.

53.6Qwen3.8 Max—Beam 1.0 501B-A23B FP8

Not comparable · BenchAlign v5.8

The coding row is not comparable on the public lane: at least one model is not measured or not ranked there.

Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.

Open the coding leaderboardShortlist models for this job
Same basis rules as the category table below

What is actually comparable

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

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.

Bars run 0–100 on each benchmark’s normalized display scale

Category results, on a stated basis

Each row shows the public-lane category score for both models: the BenchAlign v5.8 lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.

Instruction following

Like-for-like
Qwen3.8 Max
94.7
#2/148
Beam 1.0 501B-A23B FP8
89.8
#28/148
Basis
Provisional lane · 1 vs 1 weighted rows
Reading
Qwen3.8 Max leads

Agentic

Not comparable
Qwen3.8 Max
67.5
Supported · #14/128
Beam 1.0 501B-A23B FP8
Not ranked
Basis
BenchAlign v5.8 lane · 18 vs 6 public rows
Reading
Not comparable

Coding

Not comparable
Qwen3.8 Max
53.6
Supported · #34/150
Beam 1.0 501B-A23B FP8
Not ranked
Basis
BenchAlign v5.8 lane · 13 vs 7 public rows
Reading
Not comparable

Reasoning

Not comparable
Qwen3.8 Max
87.8
Unranked · 2 rankable rows
Beam 1.0 501B-A23B FP8
85.8
Unranked · 3 rankable rows
Basis
Provisional lane · 2 vs 1 weighted rows
Reading
Not comparable

Multimodal

Not comparable
Qwen3.8 Max
91.2
#4/54
Beam 1.0 501B-A23B FP8
Not ranked
Basis
Provisional lane · 2 vs 0 weighted rows
Reading
Not comparable

Knowledge

Not comparable
Qwen3.8 Max
68.1
Supported · #24/184
Beam 1.0 501B-A23B FP8
Not ranked
Basis
BenchAlign v5.8 lane · 6 vs 2 public rows
Reading
Not comparable

Multilingual

Not comparable
Qwen3.8 Max
Not ranked
Beam 1.0 501B-A23B FP8
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
Not comparable

Math

Not comparable
Qwen3.8 Max
78.9
Unranked · 2 rankable rows
Beam 1.0 501B-A23B FP8
87.9
Unranked · 1 rankable row
Basis
Provisional lane · 2 vs 1 weighted rows
Reading
Not comparable

Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.8) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.

Supported evidence per lane · bars run 0–100Methodology

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
Beam 1.0 501B-A23B FP8
API rate not published
Fits in one request

Qwen3.8 Max has no comparable published API token rate. Beam 1.0 501B-A23B FP8 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
Beam 1.0 501B-A23B FP8
API rate not published
Fits in one request

Qwen3.8 Max has no comparable published API token rate. Beam 1.0 501B-A23B FP8 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
Beam 1.0 501B-A23B FP8
API rate not published
Fits in one request
Cached-input rate unavailable

Qwen3.8 Max has no comparable published API token rate. Beam 1.0 501B-A23B FP8 has no comparable published API token rate.

Cached input falls back to the list input rate only where a cached rate is unpublished

Specification differences

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

Reasoning profile

Qwen3.8 Max

Reasoning

Beam 1.0 501B-A23B FP8

Reasoning

Weight access

Qwen3.8 Max

Open Weight

Beam 1.0 501B-A23B FP8

Open Weight

License

Qwen3.8 Max

Open Weight

Beam 1.0 501B-A23B FP8

Open Weight

Release date

Qwen3.8 Max

2026-08-03

Beam 1.0 501B-A23B FP8

2026-10-16

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.

Questions

Which is better, Qwen3.8 Max or Beam 1.0 501B-A23B FP8?

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 Beam 1.0 501B-A23B FP8?

Beam 1.0 501B-A23B FP8 is not ranked on the public lane for coding, so no winner is named for coding.

Which is better for agentic tasks, Qwen3.8 Max or Beam 1.0 501B-A23B FP8?

Beam 1.0 501B-A23B FP8 is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.

Which costs less, Qwen3.8 Max or Beam 1.0 501B-A23B FP8?

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 Beam 1.0 501B-A23B FP8?

Both models list the same context window, 1M.

Benchmark evidence

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

Browse raw public benchmark evidence76 rows

Agentic

  • Terminal-Bench 2.1

    Qwen3.8 Max86.6%
    Source
    Beam 1.0 501B-A23B FP878.9%
    Source

    Qwen3.8 Max leads this result

  • CoWorkBench

    Qwen3.8 Max74.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • JobBench

    Qwen3.8 Max53.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • skillsBench

    Qwen3.8 Max70.2%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Agents' Last Exam

    Qwen3.8 Max52.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • AutomationBench

    Qwen3.8 Max27.3%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Toolathlon-Verified

    Qwen3.8 Max72.5%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • WideResearch

    Qwen3.8 Max81.9%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • HLE w/ tools

    Qwen3.8 Max56.2%
    Source
    Beam 1.0 501B-A23B FP851.1%
    Source

    Qwen3.8 Max leads this result

  • OSWorld-Verified

    Qwen3.8 Max86.1%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • OSWorld 2.0

    Qwen3.8 Max19.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • WebArena-Verified

    Qwen3.8 Max66.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • AndroidWorld

    Qwen3.8 Max85.3%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MobileWorld

    Qwen3.8 Max77.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Terminal-Bench 2.1 (Vals)

    Qwen3.8 Max67.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Hermes Index

    Qwen3.8 Max36.2%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Hermes Bench

    Qwen3.8 Max70.5%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MazeBench

    Qwen3.8 Max0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • τ³-bench results

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP836.8%
    Source

    Not directly comparable

  • MCP Atlas

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP881.5%
    Source

    Not directly comparable

  • BrowseComp

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP876.7%
    Source

    Not directly comparable

  • DeepSearchQA

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP879.6%
    Source

    Not directly comparable

Coding

  • Terminal-Bench 2.1

    Qwen3.8 Max86.6%
    Source
    Beam 1.0 501B-A23B FP878.9%
    Source

    Qwen3.8 Max leads this result

  • SWE-bench Pro

    Qwen3.8 Max67.7%
    Source
    Beam 1.0 501B-A23B FP865.9%
    Source

    Qwen3.8 Max leads this result

  • DeepSWE

    Qwen3.8 Max56.6%
    Source
    Beam 1.0 501B-A23B FP843.0%
    Source

    Qwen3.8 Max leads this result

  • NL2Repo

    Qwen3.8 Max55.9%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • FrontierSWE

    Qwen3.8 Max73.5%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MLS-Bench Lite

    Qwen3.8 Max41.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • PaperBench

    Qwen3.8 Max93.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • VulcanBench v3

    Qwen3.8 Max81.2%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • OpenHarmony Bench

    Qwen3.8 Max60.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • FrontierSWE v2

    Qwen3.8 Max15.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • LiveCodeBench (Vals)

    Qwen3.8 Max87.9%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • SWE-bench (Vals)

    Qwen3.8 Max85.6%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Bug Hunt Bench

    Qwen3.8 Max25.7 fixes
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • SWE-Atlas Codebase QnA

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP832.2%
    Source

    Not directly comparable

  • SWE Multilingual

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP878.1%
    Source

    Not directly comparable

  • SWE-bench Verified

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP880.2%
    Source

    Not directly comparable

  • SciCode

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP848.3%
    Source

    Not directly comparable

Reasoning

  • MRCRv2

    Qwen3.8 Max92.9%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • LongBench v2

    Qwen3.8 Max66.3%
    Source
    Beam 1.0 501B-A23B FP867.6%
    Source

    Beam 1.0 501B-A23B FP8 leads this result

  • CritPt

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP817.4%
    Source

    Not directly comparable

Multimodal

  • MMMU-Pro

    Qwen3.8 Max82.3%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MathVision

    Qwen3.8 Max95.2%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MathVision w/ Python

    Qwen3.8 Max97.7%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • BabyVision

    Qwen3.8 Max82.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • BabyVision w/ Python

    Qwen3.8 Max91.3%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • ZeroBench

    Qwen3.8 Max24.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • ZeroBench w/ Python

    Qwen3.8 Max49.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MedXpertQA (MM)

    Qwen3.8 Max80.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • ScreenSpot Pro

    Qwen3.8 Max84.5%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Vision2Web

    Qwen3.8 Max69.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • CharXiv w/o tools

    Qwen3.8 Max88.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • CharXiv

    Qwen3.8 Max93.5%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • OmniDocBench 1.5

    Qwen3.8 Max92.1%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • OCRBench V2

    Qwen3.8 Max74.2%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • CC-OCR

    Qwen3.8 Max79.6%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • RealWorldQA

    Qwen3.8 Max88.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • ERQA

    Qwen3.8 Max77.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • SimpleVQA

    Qwen3.8 Max75.0%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • PerceptionBench

    Qwen3.8 Max63.5%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • Video-MME (with subtitle)

    Qwen3.8 Max90.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • VideoMMMU

    Qwen3.8 Max88.7%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MMVU

    Qwen3.8 Max82.4%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MLVU (M-Avg)

    Qwen3.8 Max90.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • LVBench

    Qwen3.8 Max81.8%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

Knowledge

  • GPQA

    Qwen3.8 Max92.6%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • GPQA-D

    Qwen3.8 Max92.6%
    Source
    Beam 1.0 501B-A23B FP890.3%
    Source

    Qwen3.8 Max leads this result

  • HLE

    Qwen3.8 Max43.6%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • HLE w/o tools

    Qwen3.8 Max43.6%
    Source
    Beam 1.0 501B-A23B FP836%
    Source

    Qwen3.8 Max leads this result

  • GPQA Diamond (Vals)

    Qwen3.8 Max93.7%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • MMLU-Pro (Vals)

    Qwen3.8 Max88.6%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

Instruction following

  • IFBench

    Qwen3.8 Max82.8%
    Source
    Beam 1.0 501B-A23B FP880.7%
    Source

    Qwen3.8 Max leads this result

Math

  • FrontierMath v2 (Tiers 1-3)

    Qwen3.8 Max74.737%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • FrontierMath v2 (Tier 4)

    Qwen3.8 Max46.341%
    Source
    Beam 1.0 501B-A23B FP8—

    Not directly comparable

  • AIME26

    Qwen3.8 Max—
    Beam 1.0 501B-A23B FP898.0%
    Source

    Not directly comparable

76 public results · 9 shared

Related comparisons

Compare API pricingRead the methodologyOpen the model selector

Watch Qwen3.8 Max vs Beam 1.0 501B-A23B FP8

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

Read a sample issue

Join 5,600+ readers.

Last updated October 16, 2026