Agents Directory
SkillsRankingsAgents
CategoriesModelsBenchmarksCompareAgent LeaderboardSkillsRankingsAgentsAbout
/Rankings
/Best Qwen models

Best Qwen models

First-party

Alibaba's Qwen family is the broadest model lineup in open AI: frontier flagships, coding specialists, vision models, and tiny MoEs that run on a laptop. These are the best Qwen models in 2026 and what each one is actually for.

The best Qwen model overall is QwenQwen3.7 Max, Alibaba's flagship and the highest-scoring Chinese model on the Artificial Analysis Intelligence Index logoArtificial Analysis Intelligence Index that we track (56.6). It is API only, at $0.80 per million input tokens.

For coding, the Qwen3 Coder line is the reason most people come to Qwen: Qwen3 Coder 480B is the open-weight specialist with a 1M token context window and a free tier on OpenRouter, QwenQwen3 Coder Next is the newer efficient line at $0.11 per million, and Qwen3 Coder 30B is the one to run locally.

For everyday volume, QwenQwen3.7 Plus ($0.40, 1M context) and Qwen3.5 Flash ($0.065, 1M context) cover the workhorse and budget tiers.

The ranking
Updated June 2026
Best Qwen flagships

The top of the lineup: maximum reasoning hosted, or the 1M-context workhorse.

#ModelContextInput
  • 1
    1Qwen
    Qwen3.7 MaxQwenThe flagship: Qwen's strongest reasoning, API only.
    Context262K
    Input$0.8/M
  • 2
    2Qwen
    Qwen3.7 PlusQwenThe everyday workhorse: 1M context at $0.40 per million input tokens.
    Context1M
    Input$0.32/M
Best Qwen for coding

The Coder line: open-weight specialists for repositories and agentic edits.

#ModelContextInput
  • 1
    1Qwen
    Qwen3 Coder 480B A35BQwenThe open-weight coding specialist: 1M context, free tier on OpenRouter.
    Context1.049M
    Input$0.22/M
  • 2
    2Qwen
    Qwen3 Coder NextQwenThe newer open coding line: efficient agentic edits at $0.11 per million.
    Context262K
    Input$0.11/M
  • 3
    3Qwen
    Qwen3 Coder 30B A3B InstructQwenThe local pick: a fast MoE coder that fits prosumer hardware.
    Context160K
    Input$0.07/M
Best value and specialty Qwen

High-volume, chat, and vision picks from the wider family.

#ModelContextInput
  • 1
    1Qwen
    Qwen3.5-FlashQwen1M context at $0.065 per million, for high-volume and routine steps.
    Context1M
    Input$0.065/M
  • 2
    2Qwen
    Qwen3 Next 80B A3B InstructQwenFast open-weight MoE with a free tier, great for chat and reasoning.
    Context262K
    Input$0.09/M
  • 3
    3Qwen
    Qwen3 VL 235B A22B InstructQwenThe open vision flagship: images and documents at $0.20 per million.
    Context262K
    Input$0.2/M
Frequently asked questions
What is the best Qwen model in 2026?

QwenQwen3.7 Max is the best Qwen model overall, with the highest Intelligence Index score (56.6) of any Chinese model we track. For coding, Qwen3 Coder 480B is the open-weight specialist most people want, and QwenQwen3.7 Plus is the everyday workhorse with a 1M token context window at $0.40 per million input tokens.

What is the best Qwen model for coding?

QwenQwen3 Coder 480B A35B is the best Qwen coding model: an open-weight mixture-of-experts specialist with a 1M token context window for large repositories, and a free tier on OpenRouter to try it. QwenQwen3 Coder Next is the newer, cheaper line for agentic edits, and Qwen3 Coder 30B is the version that fits on local hardware.

Which Qwen models are open source?

Most of the family publishes open weights, including the entire Coder line, Qwen3 Next 80B, the Qwen3 VL vision models, and the small dense models. The exceptions are the hosted flagships: QwenQwen3.7 Max, QwenQwen3.7 Plus, and the Flash tier are API only.

Can you run Qwen models locally?

Yes. Qwen3 Coder 30B A3B is the standout local pick: a mixture-of-experts coder with only 3B active parameters that runs fast on a 24GB GPU or a high-RAM Mac. QwenQwen3 32B and the 30B A3B instruct models are solid local all-rounders too. The 480B coder needs server-class hardware.

Share:
Details:
  • Models


    8
  • Filter


    First-party
  • Updated


    June 2026
Browse:SkillsRankingsModelsBenchmarksProvidersAgentsAgent LeaderboardCompareCategories
Quick Links:AboutBlog

© 2026 Agents Directory