Skip to main content

Model profile

Qwen3.6-35B-A3B

AlibabaCurrentReleased Apr 15, 2026
Data verified
Overall Score
59Prov. #45 of 78Verified #31 of 32
Arena Elo
Not listed
Categories Ranked
3of 8
Price (1M tokens)
Not listedAPI pricing
Speed
Not listed
Context
262K

Evidence coverage

58 of 300 tracked benchmarks are published. 41 are verified and 17 provisional. 7 of 8 categories are measured.

Updated July 15, 2026Methodology
Published / tracked
58 / 300
Verified
41
Provisional
17
Categories measured
7 / 8

Evidence by category

  • Agentic15 benchmarks
    Mixed evidence
  • Coding9 benchmarks
    Mixed evidence
  • Reasoning2 benchmarks
    Reported
  • Knowledge11 benchmarks
    Mixed evidence
  • Math5 benchmarks
    Verified
  • Multilingual0 benchmarks
    Not measured
  • Multimodal15 benchmarks
    Mixed evidence
  • Inst. Following1 benchmark
    Reported
Open WeightSelf-hostReasoning
Confidence:
High
base

According to BenchLM.ai, Qwen3.6-35B-A3B ranks #45 out of 78 models on the provisional leaderboard with an overall score of 59/100. It also ranks #31 out of 32 on the verified leaderboard. While not a frontier model, it offers specific advantages depending on the use case.

Qwen3.6-35B-A3B is a open weight model with a 262K token context window. It uses explicit chain-of-thought reasoning, which typically improves performance on math and complex reasoning tasks at the cost of higher latency and token usage.

This profile currently has 58 of 300 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Coding (#27), while its weakest is Knowledge (#49). This performance profile makes it particularly well-suited for software development and code generation tasks.

Peer position

Exact provisional scores and ranks for the closest listed peers.

  1. Claude Sonnet 4.5
    Anthropic
    #4359.0
    Claude Sonnet 4.5 is #43 with a score of 59.0.
    Compare
  2. Qwen3.5 397B
    Alibaba
    #4459.0
    Qwen3.5 397B is #44 with a score of 59.0.
    Compare
  3. Qwen3.6-35B-A3BCurrent model
    Alibaba
    #4559.0
    Qwen3.6-35B-A3B is #45 with a score of 59.0.
  4. Qwen3.5-122B-A10B
    Alibaba
    #4659.0
    Qwen3.5-122B-A10B is #46 with a score of 59.0.
    Compare
  5. Gemini 2.5 Pro
    Google
    #4759.0
    Gemini 2.5 Pro is #47 with a score of 59.0.
    Compare
  6. Qwen3.5-27B
    Alibaba
    #4859.0
    Qwen3.5-27B is #48 with a score of 59.0.
    Compare
  7. GLM-5V-Turbo
    Z.AI
    Unranked59.0
    GLM-5V-Turbo is Unranked with a score of 59.0.
    Compare

Category percentile

More

Relative position among models eligible for each sourced category. A higher percentile means a stronger position within that category's ranked cohort; 100 is highest.

  1. Coding72%
    Eligible cohort rank #27 of 93Category score 71.8
  2. Multimodal56%
    Eligible cohort rank #49 of 111Category score 61.8
  3. Knowledge55%
    Eligible cohort rank #49 of 107Category score 59.5

Category evidence

Scores and ranks appear only where this model has published benchmark evidence. Categories without displayable source records remain not measured.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank Not rankedWeight 22%15 benchmarksMixed sources44.7
CodingRank #27 of 93Percentile 72ndWeight 20%9 benchmarksMixed sources71.8
ReasoningRank Not rankedWeight 17%2 benchmarksReported0.0
KnowledgeRank #49 of 107Percentile 55thWeight 12%11 benchmarksMixed sources59.5
MathRank Not rankedWeight 5%5 benchmarksVerified71.8
MultilingualWeight 7%0 benchmarksNot measuredNot measured
MultimodalRank #49 of 111Percentile 56thWeight 12%15 benchmarksMixed sources61.8
Inst. FollowingRank Not rankedWeight 5%1 benchmarkReported0.0

Benchmark Details

Rows below have a displayable published verification record. Each source link and provenance note remains in the page HTML while its category is closed. Source-unverified manual rows and generated rows stay hidden.

Agentic15 benchmarks
Terminal-Bench 2.0Provider exact
51.5%Weighted 38%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
Claw-EvalProvider exact
68.7%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
QwenClawBenchProvider exact
52.6%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
QwenWebBenchProvider exact
1397Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
τ³-bench resultsProvider exact

τ³-Bench Tool-Agent-User Evaluation

67.2%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
VITA-BenchProvider exact
35.6%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
DeepPlanningProvider exact
25.9%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
ToolathlonProvider exact
26.9%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
MCP AtlasProvider exact
62.8%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
WideResearchProvider exact
60.1%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
AA Agentic IndexReported

Artificial Analysis Agentic Index

21.4%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
τ²-bench resultsReported

τ²-Bench Tool-Agent-User Evaluation

95.3%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
GDPval-AAReported

GDPval-AA normalized

27.4%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
GDPval-AAReported
1049Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Gert LabsBenchmark exact

Gert Labs Composite Game Benchmark

42.65%Display only
Source: Gert Labs rankingsProvenance: Gert Labs reports this composite leaderboard score in the public rankings API. BenchLM scales the source gscore from 0-1 to 0-100 and stores it as a display-only agentic benchmark.
Coding9 benchmarks
LiveCodeBenchProvider exact

LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code

80.4%Weighted 38%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
SWE-bench VerifiedProvider exact

Software Engineering Benchmark Verified

73.4%Weighted 16%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
SWE-bench ProProvider exact
49.5%Weighted 10%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
SWE MultilingualProvider exact
67.2%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
Terminal-Bench 2.0Provider exact
51.5%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
NL2RepoProvider exact
29.4%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
AA Coding IndexReported

Artificial Analysis Coding Index

41.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Terminal-Bench HardReported
34.8%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-SciCodeReported

Artificial Analysis SciCode

35.8%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Reasoning2 benchmarks
AA-LCRReported

Artificial Analysis Long Context Reasoning

63.7%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
CritPtReported

Critical Physics Tasks

0.3%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Knowledge11 benchmarks
HLEProvider exact

Humanity's Last Exam

21.4%Weighted 32%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
MMLU-ProProvider exact

Massive Multitask Language Understanding Professional

85.2%Weighted 22%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
SuperGPQAProvider exact

SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines

64.7%Weighted 5%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
GPQAProvider exact

Graduate-Level Google-Proof Q&A

86%Weighted 5%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
C-EvalProvider exact
90%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
Artificial Analysis Intelligence IndexReported
31.6%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-GPQA DiamondReported

Artificial Analysis GPQA Diamond

84.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-HLEReported

Artificial Analysis Humanity's Last Exam

20.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience IndexReported

Artificial Analysis Omniscience Index

-21.4%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience AccuracyReported

Artificial Analysis Omniscience Accuracy

18.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience Hallucination RateReported

Artificial Analysis Omniscience Hallucination Rate

49.7%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Math5 benchmarks
HMMT Feb 2026Provider exact

Harvard-MIT Mathematics Tournament February 2026

83.6%Weighted 25%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
AIME26Provider exact

AIME 2026

92.7%Weighted 25%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
HMMT Feb 2025Provider exact

Harvard-MIT Mathematics Tournament February 2025

90.7%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
HMMT Nov 2025Provider exact

Harvard-MIT Mathematics Tournament November 2025

89.1%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
MMAnswerBenchProvider exact
78.9%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
Multimodal15 benchmarks
MMMU-ProProvider exact

Massive Multi-discipline Multimodal Understanding Pro

75.3%Weighted 45%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
CharXivProvider exact

CharXiv Reasoning

78%Weighted 25%
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
MMMUProvider exact

Massive Multi-discipline Multimodal Understanding

81.7%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
RealWorldQAProvider exact
85.3%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
OmniDocBench 1.5Provider exact
89.9%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
SimpleVQAProvider exact
58.9%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
CC-OCRProvider exact
81.9%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
AI2D_TESTProvider exact

AI2D test split

92.7%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
RefCOCO (avg)Provider exact

RefCOCO average

92.0%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
ODINW13Provider exact
50.8%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
Video-MME (with subtitle)Provider exact

Video-MME with subtitle

86.6%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
Video-MME (w/o subtitle)Provider exact

Video-MME without subtitle

82.5%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
VideoMMMUProvider exact
83.7%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
MLVU (M-Avg)Provider exact

MLVU mean average

86.2%Display only
Source: Qwen3.6-35B-A3B model cardProvenance: Provider exact
AA-MMMU-ProReported

Artificial Analysis MMMU-Pro

75.0%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Inst. Following1 benchmark
AA-IFBenchReported

Artificial Analysis IFBench

64.4%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.

Frequently Asked Questions

How does Qwen3.6-35B-A3B perform overall in AI benchmarks?

Qwen3.6-35B-A3B currently ranks #45 out of 78 models on BenchLM's provisional leaderboard with an overall score of 59. It also ranks #31 out of 32 on the verified leaderboard. It is created by Alibaba and features a 262K context window.

Is Qwen3.6-35B-A3B good for knowledge and understanding?

Qwen3.6-35B-A3B ranks #49 out of 78 models in knowledge and understanding benchmarks with an average score of 59.5. There are stronger options in this category.

Is Qwen3.6-35B-A3B good for coding and programming?

Qwen3.6-35B-A3B ranks #27 out of 78 models in coding and programming benchmarks with an average score of 71.8. There are stronger options in this category.

Is Qwen3.6-35B-A3B good for mathematics?

Qwen3.6-35B-A3B has visible benchmark coverage in mathematics, but BenchLM does not currently assign it a global category rank there.

Is Qwen3.6-35B-A3B good for reasoning and logic?

Qwen3.6-35B-A3B has visible benchmark coverage in reasoning and logic, but BenchLM does not currently assign it a global category rank there.

Is Qwen3.6-35B-A3B good for agentic tool use and computer tasks?

Qwen3.6-35B-A3B has visible benchmark coverage in agentic tool use and computer tasks, but BenchLM does not currently assign it a global category rank there.

Is Qwen3.6-35B-A3B good for multimodal and grounded tasks?

Qwen3.6-35B-A3B ranks #49 out of 78 models in multimodal and grounded tasks benchmarks with an average score of 61.8. There are stronger options in this category.

Is Qwen3.6-35B-A3B good for instruction following?

Qwen3.6-35B-A3B has visible benchmark coverage in instruction following, but BenchLM does not currently assign it a global category rank there.

Is Qwen3.6-35B-A3B open source?

Yes, Qwen3.6-35B-A3B is an open weight model created by Alibaba, meaning it can be downloaded and run locally or fine-tuned for specific use cases.

Does Qwen3.6-35B-A3B have full benchmark coverage on BenchLM?

Not yet. Qwen3.6-35B-A3B currently has 58 published benchmark scores out of the 300 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of Qwen3.6-35B-A3B?

Qwen3.6-35B-A3B has a context window of 262K, which determines how much text it can process in a single interaction.

Last updated: July 15, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Don't miss the next GPT moment

Which models moved up, what is new, and what it costs. One email each week.

Free. One email per week.