VENICE ATLAS · MODEL DOSSIER · QWEN 3.5 35B A3B

Model overview

Qwen 3.5 35B A3B

Name: Qwen 3.5 35B A3B
Brand: venice.ai
Price: 0.1500 USD
Availability: InStock

Qwen 3.5 35B A3B is a highly efficient MoE model with 35B total parameters and only 3B active parameters. It surpasses the larger Qwen3-235B-A22B while being 6.7x smaller, excelling at reasoning, coding, and general knowledge tasks.

qwen3-5-35b-a3b · venice.ai

Text · Venice catalog$0.15 per 1M tokensPrice on Venice AI

ModalityTextOn Venice AI

Primary price$0.15 per 1M tokensper 1M tokens

Context window256KMax tokens

Benchmarks81Benchmark results

Benchmark results

See the tests behind each score. They use the benchmark’s original scoring scale.

Chat#624.67 rating Economics#914.86 rating Finance#931.11 rating Grounding#1113.69 rating Healthcare#1331.99 rating Image To Text#720.57 rating Instruction Following#1221.32 rating Language#1529.44 rating Legal#1031.11 rating Long Context#2018.65 rating Search#2911.46 rating Spatial Reasoning#1824.19 rating Structured Output#824.89 rating Text To Image#413.04 rating Video understanding#721.36 rating

Agents3 results

DeepPlanningdeep-planning · agents

22.8%

ReportedOct 2026

t2-bencht2-bench · agents

81.2%

ReportedOct 2026

VITA-Benchvita-bench · agents

31.9%

ReportedOct 2026

Chat2 results

IFEvalifeval · chat

91.9%

ReportedOct 2026

Multi-Challengemultichallenge · chat

60.0%

ReportedOct 2026

Coding2 results

FullStackBench enfullstackbench-en · code

58.1%

ReportedOct 2026

FullStackBench zhfullstackbench-zh · code

55.0%

ReportedOct 2026

General intelligence5 results

AndroidWorld_SRandroidworld-sr · general

71.1%

ReportedOct 2026

C-Evalc-eval · general

90.2%

ReportedOct 2026

Includeinclude · general

79.7%

ReportedOct 2026

MAXIFEmaxife · general

86.6%

ReportedOct 2026

NOVA-63nova-63 · general

57.1%

ReportedOct 2026

Healthcare2 results

MedXpertQAmedxpertqa · healthcare

61.4%

ReportedOct 2026

PMC-VQApmc-vqa · healthcare

62.0%

ReportedOct 2026

Instruction Following1 result

IFBenchifbench · instruction_following

70.2%

ReportedOct 2026

Language5 results

MMLU-Prommlu-pro · language

85.3%

ReportedOct 2026

MMLU-ProXmmlu-prox · language

81.0%

ReportedOct 2026

MMLU-Reduxmmlu-redux · language

93.3%

ReportedOct 2026

MMMLUmmmlu · language

85.2%

ReportedOct 2026

WMT24++wmt24++ · language

76.3%

ReportedOct 2026

Long Context1 result

LongBench v2longbench-v2 · long_context

59.0%

ReportedOct 2026

Math7 results

CodeForcescodeforces · math

82.2%

ReportedOct 2026

DynaMathdynamath · math

85.0%

ReportedOct 2026

HMMT 2025hmmt-2025 · math

89.0%

ReportedOct 2026

HMMT25hmmt25 · math

89.2%

ReportedOct 2026

MathVisionmathvision · math

83.9%

ReportedOct 2026

MathVista-Minimathvista-mini · math

86.2%

ReportedOct 2026

PolyMATHpolymath · math

64.4%

ReportedOct 2026

Multimodal2 results

MMMUmmmu · multimodal

81.4%

ReportedOct 2026

OSWorld-Verifiedosworld-verified · multimodal

54.5%

ReportedOct 2026

Learn what each benchmark measures ↗

Pricing detail

Input price per 1M tokens: $0.15
Output price per 1M tokens: $1.00
Cache read price per 1M tokens: $0.0500
Blended price per 1M tokens (50/50): $0.57

Prices are published by Venice AI. Venice model page ↗

Capabilities

ReasoningFunction callingVisionVideo inputWeb search

Endpoint metadata

Privacy: private
Quantization: not-available
Maximum completion tokens: 16K
Uncensored: No
Beta: No
Mapping confidence: 100.0%

View model source ↗

All capabilities

Max Images: 10
Quantization: not-available
Supports E2 E E: No
Supports Vision: Yes
Supports X Search: No
Optimized For Code: Yes
Supports Log Probs: Yes
Supports Reasoning: Yes
Supports Web Search: Yes
Supports Audio Input: No
Supports Video Input: Yes
Default Reasoning Effort: low
Reasoning Effort Options: none, low, medium, high
Supports Multiple Images: Yes
Supports Response Schema: Yes
Supports Tee Attestation: No
Supports Function Calling: Yes
Supports Reasoning Effort: Yes

Related models

Qwen 3.5 35B A3B

Benchmark results

More from qwen