geekradar
Int
Explore GeekRadar

More to explore

Good gear All topics About us Say hello
GeekRadar / The comparison desk
COMPARE.
AI models
alibaba

Qwen3.5 35B-A3B

AI model Released

Context window
262,144 tokens
Maximum output
65,536 tokens
Open weights
Yes

Capability & preference

Human preference · Arena

Overall text preferences from head-to-head votes. Not factual accuracy or a task-success rate.

Qwen3.5 35B-A3B 1394.5 Qwen3.5-35B-A3B · 95% interval 1390.3–1398.7 · 30,427 votes · 2026-10-28

Benchmark results

  • AI2D_TEST · version unreported · reported score92.60*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • AIME26 · version unreported · reported score · Full AIME 2026 I and II91*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Task set
    Full AIME 2026 I and II*
  • C-Eval · version unreported · reported score90.20*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • CC-OCR · version unreported · reported score80.70*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • CharXiv(RQ) · version unreported · reported score77.50*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • Claw-Eval Avg · version unreported · reported score65.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • Claw-Eval Pass^3 · version unreported · reported score51*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • DeepPlanning · version unreported · reported score22.80*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • EmbSpatialBench · version unreported · reported score83.10*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • GPQA · version unreported · reported score84.20*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • HallusionBench · version unreported · reported score67.90*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • HLE · version unreported · reported score22.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • HMMT Feb 25 · version unreported · reported score89*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • HMMT Feb 26 · version unreported · reported score78.70*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • HMMT Nov 25 · version unreported · reported score89.20*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • IMOAnswerBench · version unreported · reported score76.80*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • LiveCodeBench v6 · 6 · reported score74.60*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • LVBench · version unreported · reported score71.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • Mathvista(mini) · version unreported · reported score86.20*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • MCP-Atlas · version unreported · reported score · Public set62.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Judge
    gemini-2.5-pro*
    Task set
    Public set*
  • MCPMark · version unreported · reported score · GitHub MCP v0.30.327*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Harness
    GitHub MCP v0.30.3*
    Truncation
    Playwright responses at 32K tokens*
  • MLVU · version unreported · reported score85.60*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • MMBench EN-DEV-v1.1 · 1.1 · reported score91.50*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • MMLU-Pro · version unreported · reported score85.30*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • MMLU-Redux · version unreported · reported score93.30*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • MMMU · version unreported · reported score81.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • MMMU-Pro · version unreported · reported score75.10*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • MVBench · version unreported · reported score74.80*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • NL2Repo · version unreported · reported score · Claude Code20.50*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Harness
    Claude Code*
    Variant
    Maximum 900 turns*
    Temperature
    1.0*
    Top p
    0.95*
  • ODInW13 · version unreported · reported score42.60*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • OmniDocBench1.5 · 1.5 · reported score89.30*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • QwenClawBench · version unreported · reported score · Internal real-user distribution47.70*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Task set
    Internal real-user distribution*
    Temperature
    0.6*
    Context length
    256K tokens*
  • QwenWebBench · version unreported · reported score · Internal978* EloPlatform unreported Compare results
    Test conditions
    Metric
    BT/Elo rating*
    Checkpoint version
    Not specified*
    Languages
    English and Chinese*
    Categories
    Seven frontend application categories*
    Judge
    Multimodal automatic judge*
    Task set
    Internal*
  • RealWorldQA · version unreported · reported score84.10*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • RefCOCO(avg) · version unreported · reported score89.20*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • RefSpatialBench · version unreported · reported score63.50*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • SimpleVQA · version unreported · reported score58.30*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • SkillsBench Avg5 · version unreported · reported score · OpenCode · Self-contained subset, API-dependent tasks excluded · Mean of five runs4.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Harness
    OpenCode*
    Tasks
    78*
    Runs
    5*
    Statistic
    Mean of five runs*
    Task set
    Self-contained subset, API-dependent tasks excluded*
  • SuperGPQA · version unreported · reported score63.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • SWE-bench Multilingual · version unreported · reported score · Internal agent scaffold: bash and file editing · Public benchmark60.30*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Harness
    Internal agent scaffold: bash and file editing*
    Temperature
    1.0*
    Top p
    0.95*
    Context length
    200K tokens*
    Task set
    Public benchmark*
  • SWE-bench Pro · version unreported · reported score · Internal agent scaffold: bash and file editing · Corrected public SWE-bench Pro task set44.60*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Harness
    Internal agent scaffold: bash and file editing*
    Temperature
    1.0*
    Top p
    0.95*
    Context length
    200K tokens*
    Task set
    Corrected public SWE-bench Pro task set*
  • SWE-bench Verified · version unreported · reported score · Internal agent scaffold: bash and file editing · Public benchmark70*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Harness
    Internal agent scaffold: bash and file editing*
    Temperature
    1.0*
    Top p
    0.95*
    Context length
    200K tokens*
    Task set
    Public benchmark*
  • TAU3-Bench · version unreported · reported score68.90*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    User model
    gpt-5.2, low reasoning effort*
    Retrieval
    Default BM25*
  • Terminal-Bench 2.0 · 2.0 · reported score · Harbor/Terminus-2 · Mean of five runs40.50*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Harness
    Harbor/Terminus-2*
    Timeout
    3 hours*
    Hardware
    32 CPU / 48 GB RAM*
    Runs
    5*
    Statistic
    Mean of five runs*
    Temperature
    1.0*
    Top p
    0.95*
    Top k
    20*
    Maximum output
    80K tokens*
    Context length
    256K tokens*
  • Tool Decathlon · version unreported · reported score28.70*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • VideoMME (w sub.) · version unreported · reported score86.60*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Variant
    Subtitles enabled*
  • VideoMME (w/o sub.) · version unreported · reported score82.50*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Variant
    Subtitles disabled*
  • VideoMMMU · version unreported · reported score80.40*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • VITA-Bench · version unreported · reported score · Average subdomain score29.10*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
    Judge
    claude-4-sonnet*
    Statistic
    Average subdomain score*
  • WideSearch · version unreported · reported score59.10*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*
  • ZEROBench_sub · version unreported · reported score34.10*Platform unreported Compare results
    Test conditions
    Metric
    reported score*
    Checkpoint version
    Not specified*

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

Arena text preference / chinesehuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,456.13 rating

Confidence interval: 1,443.11 – 1,469.14

2085 votes

Category
chinese
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,409.70 rating

Confidence interval: 1,402.68 – 1,416.71

8382 votes

Category
coding
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,347.45 rating

Confidence interval: 1,338.51 – 1,356.40

4672 votes

Category
creative_writing
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,406.71 rating

Confidence interval: 1,400.97 – 1,412.45

13554 votes

Category
english
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,373.14 rating

Confidence interval: 1,367.38 – 1,378.89

22179 votes

Category
exclude_ties
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,407.14 rating

Confidence interval: 1,395.94 – 1,418.34

2795 votes

Category
expert
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,413.45 rating

Confidence interval: 1,392.63 – 1,434.27

881 votes

Category
french
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,368.57 rating

Confidence interval: 1,343.28 – 1,393.86

486 votes

Category
german
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,399.73 rating

Confidence interval: 1,394.62 – 1,404.85

19316 votes

Category
hard_prompts
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,410.36 rating

Confidence interval: 1,403.51 – 1,417.21

8835 votes

Category
hard_prompts_english
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,393.17 rating

Confidence interval: 1,385.34 – 1,401.00

6107 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,347.36 rating

Confidence interval: 1,339.31 – 1,355.41

5945 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,398.44 rating

Confidence interval: 1,386.11 – 1,410.78

2381 votes

Category
industry_legal_and_government
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,412.81 rating

Confidence interval: 1,404.37 – 1,421.25

5191 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,410.30 rating

Confidence interval: 1,395.97 – 1,424.64

1728 votes

Category
industry_mathematical
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,421.11 rating

Confidence interval: 1,408.48 – 1,433.74

2273 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,414.96 rating

Confidence interval: 1,408.94 – 1,420.98

12042 votes

Category
industry_software_and_it_services
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,365.16 rating

Confidence interval: 1,357.74 – 1,372.57

7047 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,379.25 rating

Confidence interval: 1,372.81 – 1,385.69

9923 votes

Category
instruction_following
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,325.43 rating

Confidence interval: 1,286.71 – 1,364.15

251 votes

Category
japanese
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,353.39 rating

Confidence interval: 1,326.00 – 1,380.78

483 votes

Category
korean
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,389.16 rating

Confidence interval: 1,382.92 – 1,395.41

11895 votes

Category
longer_query
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,403.91 rating

Confidence interval: 1,390.27 – 1,417.55

1851 votes

Category
math
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,390.17 rating

Confidence interval: 1,381.79 – 1,398.56

5476 votes

Category
multi_turn
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,378.10 rating

Confidence interval: 1,372.87 – 1,383.32

16873 votes

Category
non_english
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,394.54 rating

Confidence interval: 1,390.34 – 1,398.74

30427 votes

Category
overall
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,385.19 rating

Confidence interval: 1,362.22 – 1,408.17

586 votes

Category
polish
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,376.61 rating

Confidence interval: 1,366.62 – 1,386.59

3629 votes

Category
russian
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-28 · rating

Qwen3.5 35B-A3B

1,389.67 rating

Confidence interval: 1,370.50 – 1,408.83

1005 votes

Category
spanish
Evaluated configuration
Qwen3.5-35B-A3B
Evaluation type
human preference
Arena text preference / chinesehuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,456.60 rating

Confidence interval: 1,443.62 – 1,469.59

2090 votes

Category
chinese
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,409.83 rating

Confidence interval: 1,402.82 – 1,416.84

8395 votes

Category
coding
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,347.61 rating

Confidence interval: 1,338.67 – 1,356.55

4680 votes

Category
creative_writing
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,406.87 rating

Confidence interval: 1,401.13 – 1,412.61

13568 votes

Category
english
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,373.41 rating

Confidence interval: 1,367.65 – 1,379.17

22203 votes

Category
exclude_ties
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,407.87 rating

Confidence interval: 1,396.67 – 1,419.07

2802 votes

Category
expert
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,411.73 rating

Confidence interval: 1,390.96 – 1,432.51

888 votes

Category
french
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,367.20 rating

Confidence interval: 1,341.87 – 1,392.53

488 votes

Category
german
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,399.86 rating

Confidence interval: 1,394.74 – 1,404.98

19334 votes

Category
hard_prompts
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,410.46 rating

Confidence interval: 1,403.60 – 1,417.31

8846 votes

Category
hard_prompts_english
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,393.58 rating

Confidence interval: 1,385.75 – 1,401.41

6113 votes

Category
industry_business_and_management_and_financial_operations
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_entertainment_and_sports_and_mediahuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,347.28 rating

Confidence interval: 1,339.23 – 1,355.33

5953 votes

Category
industry_entertainment_and_sports_and_media
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_legal_and_governmenthuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,398.85 rating

Confidence interval: 1,386.52 – 1,411.19

2384 votes

Category
industry_legal_and_government
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_life_and_physical_and_social_sciencehuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,413.11 rating

Confidence interval: 1,404.67 – 1,421.54

5196 votes

Category
industry_life_and_physical_and_social_science
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_mathematicalhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,410.32 rating

Confidence interval: 1,395.98 – 1,424.66

1727 votes

Category
industry_mathematical
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_medicine_and_healthcarehuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,421.29 rating

Confidence interval: 1,408.68 – 1,433.89

2284 votes

Category
industry_medicine_and_healthcare
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,415.15 rating

Confidence interval: 1,409.13 – 1,421.18

12054 votes

Category
industry_software_and_it_services
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,365.22 rating

Confidence interval: 1,357.80 – 1,372.63

7053 votes

Category
industry_writing_and_literature_and_language
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / instruction_followinghuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,379.47 rating

Confidence interval: 1,373.03 – 1,385.91

9934 votes

Category
instruction_following
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / japanesehuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,325.55 rating

Confidence interval: 1,286.75 – 1,364.34

251 votes

Category
japanese
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / koreanhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,354.17 rating

Confidence interval: 1,326.78 – 1,381.56

484 votes

Category
korean
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / longer_queryhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,389.57 rating

Confidence interval: 1,383.32 – 1,395.81

11908 votes

Category
longer_query
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / mathhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,404.13 rating

Confidence interval: 1,390.48 – 1,417.78

1850 votes

Category
math
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / multi_turnhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,390.17 rating

Confidence interval: 1,381.79 – 1,398.56

5481 votes

Category
multi_turn
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / non_englishhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,378.42 rating

Confidence interval: 1,373.19 – 1,383.64

16889 votes

Category
non_english
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / overallhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,394.73 rating

Confidence interval: 1,390.53 – 1,398.94

30457 votes

Category
overall
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / polishhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,385.83 rating

Confidence interval: 1,362.86 – 1,408.80

585 votes

Category
polish
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / russianhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,377.02 rating

Confidence interval: 1,367.03 – 1,387.00

3630 votes

Category
russian
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference
Arena text preference / spanishhuman preference · 2026-10-15 · rating

Qwen3.5 35B-A3B

1,390.92 rating

Confidence interval: 1,371.71 – 1,410.13

1003 votes

Category
spanish
Evaluated configuration
qwen3.5-35b-a3b
Evaluation type
human preference

Run it locally

Inference · memory estimate

Qwen3.5 35B-A3B

67.0 GiB checkpoint tensor payload · 35.95 B stored tensor elements · BF16/F32

Stored tensors can include buffers and repeated shared weights. A verified unique parameter count is unavailable. Checkpoint size is not the memory required to run the model.

API cost

USD / month

Standard text-token estimate. Excludes taxes, caching, batch discounts, reasoning and multimodal surcharges. Provider limits and regional availability may differ.

Qwen3.5 35B-A3B

Qwen3.5 35B-A3B / Pricing tiers 6
Provider / scopeRateCurrencyAmount / unit
DeepInfraQwen/Qwen3.5-35B-A3BStandard Cached inputUSD0.05 / 1M tokens
InputUSD0.14 / 1M tokens
OutputUSD1 / 1M tokens
SiliconFlowQwen/Qwen3.5-35B-A3BStandard 1314 InputCNY1.6 / 1M tokens
OutputCNY12.8 / 1M tokens
SiliconFlowQwen/Qwen3.5-35B-A3BStandard 输入 [0,128k) Token tier: 0 – 128,000 InputCNY0.4 / 1M tokens
OutputCNY3.2 / 1M tokens
SiliconFlowQwen/Qwen3.5-35B-A3BStandard 输入 [0, 128k) Token tier: 0 – 128,000 InputCNY0.4 / 1M tokens
OutputCNY3.2 / 1M tokens
SiliconFlowQwen/Qwen3.5-35B-A3BStandard 输入 [128k, +∞) Token tier: 128,000 – No stated upper limit InputCNY1.6 / 1M tokens
OutputCNY12.8 / 1M tokens
SiliconFlowQwen/Qwen3.5-35B-A3BStandard 输入 [128k,256k) Token tier: 128,000 – 256,000 InputCNY1.6 / 1M tokens
OutputCNY12.8 / 1M tokens

Capabilities

Context window
262,144 tokens
Maximum output
65,536 tokens
Reasoning
Yes
Tool calling
Yes
Open weights
Yes
Checkpoint tensor payload
66.97 GiB

Input modalities

  • text
  • image
  • video

Output modalities

  • text

Processing & platform

Architecture
qwen3_5_moe

System configuration

Detailed specifications

Specifications · 51 fields
HallusionBench
suite
HallusionBench
workload
reported score
score
67.9*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
NL2Repo
suite
NL2Repo
workload
reported score · Claude Code
score
20.5*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Harness
Claude Code*
Variant
Maximum 900 turns*
Temperature
1.0*
Top p
0.95*
QwenWebBench
suite
QwenWebBench
workload
reported score · Internal
score
978.0*
unit
Elo
Scope and conditions
Test conditions
Metric
BT/Elo rating*
Checkpoint version
Not specified*
Languages
English and Chinese*
Categories
Seven frontend application categories*
Judge
Multimodal automatic judge*
Task set
Internal*
OmniDocBench1.5
suite
OmniDocBench1.5
version
1.5
workload
reported score
score
89.3*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
CharXiv(RQ)
suite
CharXiv(RQ)
workload
reported score
score
77.5*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
TAU3-Bench
suite
TAU3-Bench
workload
reported score
score
68.9*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
User model
gpt-5.2, low reasoning effort*
Retrieval
Default BM25*
VITA-Bench
suite
VITA-Bench
workload
reported score · Average subdomain score
score
29.1*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Judge
claude-4-sonnet*
Statistic
Average subdomain score*
CC-OCR
suite
CC-OCR
workload
reported score
score
80.7*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
DeepPlanning
suite
DeepPlanning
workload
reported score
score
22.8*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
AI2D TEST
suite
AI2D_TEST
workload
reported score
score
92.6*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Tool Decathlon
suite
Tool Decathlon
workload
reported score
score
28.7*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MCPMark
suite
MCPMark
workload
reported score · GitHub MCP v0.30.3
score
27.0*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Harness
GitHub MCP v0.30.3*
Truncation
Playwright responses at 32K tokens*
RefCOCO(avg)
suite
RefCOCO(avg)
workload
reported score
score
89.2*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
ODInW13
suite
ODInW13
workload
reported score
score
42.6*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MCP-Atlas
suite
MCP-Atlas
workload
reported score · Public set
score
62.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Judge
gemini-2.5-pro*
Task set
Public set*
EmbSpatialBench
suite
EmbSpatialBench
workload
reported score
score
83.1*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
WideSearch
suite
WideSearch
workload
reported score
score
59.1*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MMMU
suite
MMMU
workload
reported score
score
81.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
SWE-bench Verified
suite
SWE-bench Verified
workload
reported score · Internal agent scaffold: bash and file editing · Public benchmark
score
70.0*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Harness
Internal agent scaffold: bash and file editing*
Temperature
1.0*
Top p
0.95*
Context length
200K tokens*
Task set
Public benchmark*
RefSpatialBench
suite
RefSpatialBench
workload
reported score
score
63.5*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MMLU-Pro
suite
MMLU-Pro
workload
reported score
score
85.3*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MMLU-Redux
suite
MMLU-Redux
workload
reported score
score
93.3*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
VideoMME (w sub.)
suite
VideoMME (w sub.)
workload
reported score
score
86.6*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Variant
Subtitles enabled*
SuperGPQA
suite
SuperGPQA
workload
reported score
score
63.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
VideoMME (w/o sub.)
suite
VideoMME (w/o sub.)
workload
reported score
score
82.5*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Variant
Subtitles disabled*
VideoMMMU
suite
VideoMMMU
workload
reported score
score
80.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
C-Eval
suite
C-Eval
workload
reported score
score
90.2*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MLVU
suite
MLVU
workload
reported score
score
85.6*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
GPQA
suite
GPQA
workload
reported score
score
84.2*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MVBench
suite
MVBench
workload
reported score
score
74.8*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
LVBench
suite
LVBench
workload
reported score
score
71.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
HLE
suite
HLE
workload
reported score
score
22.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
LiveCodeBench v6
suite
LiveCodeBench v6
version
6
workload
reported score
score
74.6*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
HMMT Feb 25
suite
HMMT Feb 25
workload
reported score
score
89.0*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
MMMU-Pro
suite
MMMU-Pro
workload
reported score
score
75.1*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
SWE-bench Multilingual
suite
SWE-bench Multilingual
workload
reported score · Internal agent scaffold: bash and file editing · Public benchmark
score
60.3*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Harness
Internal agent scaffold: bash and file editing*
Temperature
1.0*
Top p
0.95*
Context length
200K tokens*
Task set
Public benchmark*
HMMT Nov 25
suite
HMMT Nov 25
workload
reported score
score
89.2*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
HMMT Feb 26
suite
HMMT Feb 26
workload
reported score
score
78.7*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
IMOAnswerBench
suite
IMOAnswerBench
workload
reported score
score
76.8*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
AIME26
suite
AIME26
workload
reported score · Full AIME 2026 I and II
score
91.0*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Task set
Full AIME 2026 I and II*
SWE-bench Pro
suite
SWE-bench Pro
workload
reported score · Internal agent scaffold: bash and file editing · Corrected public SWE-bench Pro task set
score
44.6*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Harness
Internal agent scaffold: bash and file editing*
Temperature
1.0*
Top p
0.95*
Context length
200K tokens*
Task set
Corrected public SWE-bench Pro task set*
Mathvista(mini)
suite
Mathvista(mini)
workload
reported score
score
86.2*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
ZEROBench sub
suite
ZEROBench_sub
workload
reported score
score
34.1*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Terminal-Bench 2.0
suite
Terminal-Bench 2.0
version
2.0
workload
reported score · Harbor/Terminus-2 · Mean of five runs
score
40.5*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Harness
Harbor/Terminus-2*
Timeout
3 hours*
Hardware
32 CPU / 48 GB RAM*
Runs
5*
Statistic
Mean of five runs*
Temperature
1.0*
Top p
0.95*
Top k
20*
Maximum output
80K tokens*
Context length
256K tokens*
Claw-Eval Avg
suite
Claw-Eval Avg
workload
reported score
score
65.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
RealWorldQA
suite
RealWorldQA
workload
reported score
score
84.1*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Claw-Eval Pass^3
suite
Claw-Eval Pass^3
workload
reported score
score
51.0*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
SkillsBench Avg5
suite
SkillsBench Avg5
workload
reported score · OpenCode · Self-contained subset, API-dependent tasks excluded · Mean of five runs
score
4.4*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Harness
OpenCode*
Tasks
78*
Runs
5*
Statistic
Mean of five runs*
Task set
Self-contained subset, API-dependent tasks excluded*
MMBench EN-DEV-v1.1
suite
MMBench EN-DEV-v1.1
version
1.1
workload
reported score
score
91.5*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
SimpleVQA
suite
SimpleVQA
workload
reported score
score
58.3*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
QwenClawBench
suite
QwenClawBench
workload
reported score · Internal real-user distribution
score
47.7*
Scope and conditions
Test conditions
Metric
reported score*
Checkpoint version
Not specified*
Task set
Internal real-user distribution*
Temperature
0.6*
Context length
256K tokens*

* Provisional data