VENICE ATLAS · MODEL DOSSIER · QWEN 3.8 MAX

Model overview

Qwen 3.8 Max

Name: Qwen 3.8 Max
Brand: venice.ai
Price: 2.5000 USD
Availability: InStock

Qwen 3.8 Max is Alibaba's flagship 2.4-trillion-parameter MoE model, with major gains over Qwen 3.7 Max in software engineering and office-productivity workflows and strong long-horizon, multi-agent performance. It accepts both text and vision-language input (images and video), operates in thinking mode only, and supports a 1M-token context window.

qwen-3-8-max · venice.ai

Text · Venice catalog$2.50 per 1M tokensPrice on Venice AI

ModalityTextOn Venice AI

Primary price$2.50 per 1M tokensper 1M tokens

Context window1MMax tokens

Benchmarks42Benchmark results

Benchmark results

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

Agents#1535.69 rating Biology#1335.42 rating Chemistry#1735.42 rating Coding#1936.94 rating Vision#637.76 rating General intelligence#1350.75 rating Grounding#521.90 rating Instruction Following#131.75 rating Long Context#429.06 rating Multimodal#338.78 rating Physics#1735.42 rating Productivity#513.97 rating Reasoning#1249.85 rating Search#1523.03 rating Spatial Reasoning#232.89 rating Structured Output#527.47 rating Tool Calling#1129.32 rating Video understanding#523.73 rating

Agents12 results

Agents' Last Examagents-last-exam · agents

52.4%

ReportedOct 2026

AndroidBenchandroidbench · agents

75.1%

ReportedOct 2026

AndroidWorldandroidworld · agents

85.3%

ReportedOct 2026

AutomationBenchautomationbench · agents

27.3%

ReportedOct 2026

CoWorkBenchcoworkbench · agents

74.8%

ReportedOct 2026

Job Benchjob-bench · agents

53.4%

ReportedOct 2026

MLS-Bench Litemls-bench-lite · agents

41.0%

ReportedOct 2026

MobileWorldmobileworld · agents

77.8%

ReportedOct 2026

OneMillion Benchonemillion-bench · agents

52.5%

ReportedOct 2026

QwenSVGqwen-svg · agents

85.7%

ReportedOct 2026

Toolathlontoolathlon · agents

72.5%

ReportedOct 2026

Workspace Benchworkspace-bench · agents

67.7%

ReportedOct 2026

Coding9 results

DeepSWE 1.1deepswe-1.1 · code

56.6%

ReportedOct 2026

FrontierSWEfrontierswe · code

73.5%

ReportedOct 2026

NL2Reponl2repo · code

55.9%

ReportedOct 2026

PaperBenchpaperbench · code

93.0%

ReportedOct 2026

QwenQoderBenchqwen-qoder-bench · code

58.4%

ReportedOct 2026

QwenReactBenchqwen-react-bench · code

86.2%

ReportedOct 2026

QwenSWEBenchqwen-swe-bench · code

80.7%

ReportedOct 2026

SkillsBenchskillsbench · code

70.2%

ReportedOct 2026

Vision2Webvision2web · code

69.0%

ReportedOct 2026

Healthcare1 result

HealthBenchhealthbench · healthcare

60.2%

ReportedOct 2026

Instruction Following1 result

IFBenchifbench · instruction_following

82.8%

ReportedOct 2026

Knowledge3 results

PLawBenchplawbench · knowledge

73.2%

ReportedOct 2026

PRBench-Financeprbench-finance · knowledge

58.3%

ReportedOct 2026

PRBench-Legalprbench-legal · knowledge

57.6%

ReportedOct 2026

Long Context2 results

LongBench v2longbench-v2 · long_context

66.3%

ReportedOct 2026

MRCR v2 (8-needle)mrcr-v2-(8-needle) · long_context

92.9%

ReportedOct 2026

Multimodal1 result

OSWorld-Verifiedosworld-verified · multimodal

86.1%

ReportedOct 2026

Reasoning1 result

GPQAgpqa · reasoning

92.6%

ReportedOct 2026

Learn what each benchmark measures ↗

Pricing detail

Input price per 1M tokens: $2.50
Output price per 1M tokens: $7.50
Cache read price per 1M tokens: $0.31
Cache write price per 1M tokens: $3.13
Blended price per 1M tokens (50/50): $5.00

Prices are published by Venice AI. Venice model page ↗

Capabilities

ReasoningFunction callingVisionVideo inputWeb search

Endpoint metadata

Privacy: anonymized
Quantization: not-available
Maximum completion tokens: 131K
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: No
Supports Reasoning: Yes
Supports Web Search: Yes
Supports Audio Input: No
Supports Video Input: Yes
Supports Multiple Images: Yes
Supports Response Schema: No
Supports Tee Attestation: No
Supports Function Calling: Yes
Supports Reasoning Effort: No

Related models

Qwen 3.8 Max

Benchmark results

More from qwen