Agentic
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- 67.3
- Supported · #9/154
- Basis
- BenchAlign lane · 1 vs 15 public rows
- Reading
- Not comparable
Pareto 26.9 vs Qwen3.8 Max
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.
3 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Updated September 18, 2026. We do not rank this pair: at least one has no public score. Public scores include evidence status and uncertainty.
Both of these models will change. Get the price, version and retirement notices for the pair, each with its source. Follow model changes
Share or export
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
Pareto 26.9 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
Pareto 26.9 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
Not enough matched evidence
A complete context comparison is not sourced.
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.
1 Pareto 26.9 only
3 Shared
57 Qwen3.8 Max only
- Shared results
- 3
- Pareto 26.9 only
- 1
- Qwen3.8 Max only
- 57
- Like-for-like categories
- 0 / 8
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign 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.
Coding
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- 60.7
- Supported · #20/154
- Basis
- BenchAlign lane · 1 vs 12 public rows
- Reading
- Not comparable
Reasoning
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- 86.6
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Knowledge
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- 69.1
- Supported · #17/184
- Basis
- BenchAlign lane · 1 vs 6 public rows
- Reading
- Not comparable
Math
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- 87.4
- #5/48
- Basis
- Provisional lane · 1 vs 2 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Pareto 26.9
- Not ranked
- Qwen3.8 Max
- 90.5
- #16/124
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
| Category | Pareto 26.9 | Qwen3.8 Max | Basis | Reading |
|---|---|---|---|---|
| Agentic | Not ranked | 67.3Supported · #9/154 | Not comparableBenchAlign lane · 1 vs 15 public rows | Not comparable |
| Coding | Not ranked | 60.7Supported · #20/154 | Not comparableBenchAlign lane · 1 vs 12 public rows | Not comparable |
| Reasoning | Not ranked | 86.6Unranked · 2 rankable rows | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
| Knowledge | Not ranked | 69.1Supported · #17/184 | Not comparableBenchAlign lane · 1 vs 6 public rows | Not comparable |
| Math | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | Not ranked | 87.4#5/48 | Not comparableProvisional lane · 1 vs 2 weighted rows | Not comparable |
| Instruction following | Not ranked | 90.5#16/124 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
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.
HLE w/o tools
Knowledge
MMMU-Pro
Multimodal
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
- Pareto 26.9
- $0.00625
- Fit state unavailable
- Qwen3.8 Max
- API rate not published
- Fits in one request
Qwen3.8 Max has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- Pareto 26.9
- $0.1475
- Fit state unavailable
- Qwen3.8 Max
- API rate not published
- Fits in one request
Qwen3.8 Max has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Pareto 26.9
- $0.175
- Fit state unavailable
- Qwen3.8 Max
- API rate not published
- Fits in one request
- Cached-input rate unavailable
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.
SpecificationPareto 26.9Qwen3.8 Max
Context window
Maximum documented context; output-token limits may be lower.
Pareto 26.9
N/A
Qwen3.8 Max
API model ID
Pareto 26.9
pareto
Unbiased pricingQwen3.8 Max
qwen3.8-max
Alibaba Cloud Model Studio pricingCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Pareto 26.9
$0.25 per 1M cached input tokens
Unbiased pricingQwen3.8 Max
No comparable hosted API rate
Alibaba Cloud Model Studio pricingDocumented inputs
Pareto 26.9
Not sourced
Qwen3.8 Max
Not sourced
Documented outputs
Pareto 26.9
Not sourced
Qwen3.8 Max
Not sourced
Provider availability
Pareto 26.9
Not sourced
Qwen3.8 Max
Not sourced
Reasoning profile
Pareto 26.9
Reasoning
Qwen3.8 Max
Reasoning
Weight access
Pareto 26.9
Proprietary
Qwen3.8 Max
Open Weight
License
Pareto 26.9
Proprietary
Qwen3.8 Max
Open Weight
Release date
Pareto 26.9
2026-09-17
Qwen3.8 Max
2026-08-03
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
- A complete documented context comparison is not available.
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 evidence61 rows
Agentic
Terminal-Bench 4.0
Pareto 26.951.00%SourceQwen3.8 Max—Not directly comparable
Terminal-Bench 2.1
Pareto 26.9—Qwen3.8 Max86.6%SourceNot directly comparable
CoWorkBench
Pareto 26.9—Qwen3.8 Max74.8%SourceNot directly comparable
JobBench
Pareto 26.9—Qwen3.8 Max53.4%SourceNot directly comparable
skillsBench
Pareto 26.9—Qwen3.8 Max70.2%SourceNot directly comparable
Agents' Last Exam
Pareto 26.9—Qwen3.8 Max52.4%SourceNot directly comparable
AutomationBench
Pareto 26.9—Qwen3.8 Max27.3%SourceNot directly comparable
Toolathlon-Verified
Pareto 26.9—Qwen3.8 Max72.5%SourceNot directly comparable
WideResearch
Pareto 26.9—Qwen3.8 Max81.9%SourceNot directly comparable
HLE w/ tools
Pareto 26.9—Qwen3.8 Max56.2%SourceNot directly comparable
OSWorld-Verified
Pareto 26.9—Qwen3.8 Max86.1%SourceNot directly comparable
OSWorld 2.0
Pareto 26.9—Qwen3.8 Max19.4%SourceNot directly comparable
WebArena-Verified
Pareto 26.9—Qwen3.8 Max66.8%SourceNot directly comparable
AndroidWorld
Pareto 26.9—Qwen3.8 Max85.3%SourceNot directly comparable
MobileWorld
Pareto 26.9—Qwen3.8 Max77.8%SourceNot directly comparable
Terminal-Bench 2.1 (Vals)
Pareto 26.9—Qwen3.8 Max67.4%SourceNot directly comparable
Coding
DeepSWE
Pareto 26.974.0%SourceQwen3.8 Max56.6%SourcePareto 26.9 leads this result
Terminal-Bench 2.1
Pareto 26.9—Qwen3.8 Max86.6%SourceNot directly comparable
SWE-bench Pro
Pareto 26.9—Qwen3.8 Max67.7%SourceNot directly comparable
NL2Repo
Pareto 26.9—Qwen3.8 Max55.9%SourceNot directly comparable
FrontierSWE
Pareto 26.9—Qwen3.8 Max73.5%SourceNot directly comparable
MLS-Bench Lite
Pareto 26.9—Qwen3.8 Max41.0%SourceNot directly comparable
PaperBench
Pareto 26.9—Qwen3.8 Max93.0%SourceNot directly comparable
VulcanBench v3
Pareto 26.9—Qwen3.8 Max81.2%SourceNot directly comparable
OpenHarmony Bench
Pareto 26.9—Qwen3.8 Max60.8%SourceNot directly comparable
FrontierSWE v2
Pareto 26.9—Qwen3.8 Max15.8%SourceNot directly comparable
LiveCodeBench (Vals)
Pareto 26.9—Qwen3.8 Max87.9%SourceNot directly comparable
SWE-bench (Vals)
Pareto 26.9—Qwen3.8 Max85.6%SourceNot directly comparable
Reasoning
Knowledge
HLE w/o tools
Pareto 26.949%SourceQwen3.8 Max43.6%SourcePareto 26.9 leads this result
GPQA
Pareto 26.9—Qwen3.8 Max92.6%SourceNot directly comparable
GPQA-D
Pareto 26.9—Qwen3.8 Max92.6%SourceNot directly comparable
HLE
Pareto 26.9—Qwen3.8 Max43.6%SourceNot directly comparable
GPQA Diamond (Vals)
Pareto 26.9—Qwen3.8 Max93.7%SourceNot directly comparable
MMLU-Pro (Vals)
Pareto 26.9—Qwen3.8 Max88.6%SourceNot directly comparable
Multimodal
MMMU-Pro
Pareto 26.978%SourceQwen3.8 Max82.3%SourceQwen3.8 Max leads this result
MathVision
Pareto 26.9—Qwen3.8 Max95.2%SourceNot directly comparable
MathVision w/ Python
Pareto 26.9—Qwen3.8 Max97.7%SourceNot directly comparable
BabyVision
Pareto 26.9—Qwen3.8 Max82.0%SourceNot directly comparable
BabyVision w/ Python
Pareto 26.9—Qwen3.8 Max91.3%SourceNot directly comparable
ZeroBench
Pareto 26.9—Qwen3.8 Max24.0%SourceNot directly comparable
ZeroBench w/ Python
Pareto 26.9—Qwen3.8 Max49.0%SourceNot directly comparable
MedXpertQA (MM)
Pareto 26.9—Qwen3.8 Max80.4%SourceNot directly comparable
ScreenSpot Pro
Pareto 26.9—Qwen3.8 Max84.5%SourceNot directly comparable
Vision2Web
Pareto 26.9—Qwen3.8 Max69.0%SourceNot directly comparable
CharXiv w/o tools
Pareto 26.9—Qwen3.8 Max88.4%SourceNot directly comparable
CharXiv
Pareto 26.9—Qwen3.8 Max93.5%SourceNot directly comparable
OmniDocBench 1.5
Pareto 26.9—Qwen3.8 Max92.1%SourceNot directly comparable
OCRBench V2
Pareto 26.9—Qwen3.8 Max74.2%SourceNot directly comparable
CC-OCR
Pareto 26.9—Qwen3.8 Max79.6%SourceNot directly comparable
RealWorldQA
Pareto 26.9—Qwen3.8 Max88.0%SourceNot directly comparable
ERQA
Pareto 26.9—Qwen3.8 Max77.8%SourceNot directly comparable
SimpleVQA
Pareto 26.9—Qwen3.8 Max75.0%SourceNot directly comparable
PerceptionBench
Pareto 26.9—Qwen3.8 Max63.5%SourceNot directly comparable
Video-MME (with subtitle)
Pareto 26.9—Qwen3.8 Max90.4%SourceNot directly comparable
VideoMMMU
Pareto 26.9—Qwen3.8 Max88.7%SourceNot directly comparable
MMVU
Pareto 26.9—Qwen3.8 Max82.4%SourceNot directly comparable
MLVU (M-Avg)
Pareto 26.9—Qwen3.8 Max90.8%SourceNot directly comparable
LVBench
Pareto 26.9—Qwen3.8 Max81.8%SourceNot directly comparable
Instruction following
IFBench
Pareto 26.9—Qwen3.8 Max82.8%SourceNot directly comparable
Questions
Which is better, Pareto 26.9 or Qwen3.8 Max?
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, Pareto 26.9 or Qwen3.8 Max?
Pareto 26.9 is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, Pareto 26.9 or Qwen3.8 Max?
Pareto 26.9 is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, Pareto 26.9 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, Pareto 26.9 or Qwen3.8 Max?
A complete documented context-window comparison is not available.
Related comparisons
Last updated September 18, 2026
Watch Pareto 26.9 vs Qwen3.8 Max
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.