modelgrep

Open-Source Qwen Models

Quick answer · Updated June 2026

The best open-weight Qwen model is Qwen3.5 397B A17B (40.1 intelligence) — downloadable from Hugging Face and self-hostable. Qwen3.5-27B (37.2) and Qwen3.6 27B (37.1) round out the top three.

40.1Intelligence
134 t/sSpeed
$0.390Input /M
262KContext

The best open-weight AI models — downloadable from Hugging Face and self-hostable — ranked by intelligence. Open-source alternatives to proprietary frontier models.

  1. 1Q
    qwen3.5-397b-a17b
    ReasoningToolsJSON+140.1 intel · $0.390/M · 134 t/s
    40.1
    Intelligence
  2. 2Q
    qwen3.5-27b
    ReasoningToolsJSON+137.2 intel · $0.195/M · 35 t/s
    37.2
    Intelligence
  3. 3Q
    qwen3.6-27b
    ReasoningToolsJSON+137.1 intel · $0.288/M · 76 t/s
    37.1
    Intelligence
  4. 4Q
    qwen3.5-122b-a10b
    ReasoningToolsJSON+135.9 intel · $0.260/M · 100 t/s
    35.9
    Intelligence
  5. 5Q
    qwen3.5-9b
    ReasoningToolsJSON+132.4 intel · $0.100/M · 88 t/s
    32.4
    Intelligence
  6. 6Q
    qwen3.6-35b-a3b
    ReasoningToolsJSON+131.5 intel · $0.150/M · 173 t/s
    31.5
    Intelligence
  7. 7Q
    qwen3.5-35b-a3b
    ReasoningToolsJSON+130.7 intel · $0.140/M · 174 t/s
    30.7
    Intelligence
  8. 8Q
    qwen3-235b-a22b-thinking-2507
    ReasoningToolsJSON29.5 intel · $0.100/M · 92 t/s
    29.5
    Intelligence
  9. 9Q
    qwen3-coder-next
    ToolsJSON28.3 intel · $0.110/M · 87 t/s
    28.3
    Intelligence
  10. 10Q
    qwen3-vl-235b-a22b-thinking
    ReasoningToolsJSON+127.6 intel · $0.260/M · 45 t/s
    27.6
    Intelligence
  11. 11Q
    qwen3-next-80b-a3b-thinking
    ReasoningToolsJSON26.7 intel · $0.098/M · 168 t/s
    26.7
    Intelligence
  12. 12Q
    qwen3-235b-a22b-2507
    ToolsJSON25.0 intel · $0.090/M · 82 t/s
    25.0
    Intelligence
  13. 13Q
    qwen3-coder:free
    Tools24.8 intel · Free/M · 45 t/s
    24.8
    Intelligence
  14. 14Q
    qwen3-coder
    ToolsJSON24.8 intel · $0.220/M · 45 t/s
    24.8
    Intelligence
  15. 15Q
    qwen3-30b-a3b-thinking-2507
    ReasoningToolsJSON22.4 intel · $0.080/M · 134 t/s
    22.4
    Intelligence
  16. 16Q
    qwen3-next-80b-a3b-instruct:free
    ToolsJSON20.1 intel · Free/M · 81 t/s
    20.1
    Intelligence
  17. 17Q
    qwen3-next-80b-a3b-instruct
    ToolsJSON20.1 intel · $0.090/M · 81 t/s
    20.1
    Intelligence
  18. 18Q
    qwen3-coder-30b-a3b-instruct
    ToolsJSON20.0 intel · $0.070/M · 74 t/s
    20.0
    Intelligence
  19. 19Q
    qwen3-235b-a22b
    ReasoningToolsJSON19.8 intel · $0.455/M · 53 t/s
    19.8
    Intelligence
  20. 20Q
    qwen3-vl-30b-a3b-thinking
    ReasoningToolsJSON+119.7 intel · $0.130/M · 123 t/s
    19.7
    Intelligence
  21. 21Q
    qwen3-vl-32b-instruct
    ToolsJSONVision17.2 intel · $0.104/M · 55 t/s
    17.2
    Intelligence
  22. 22Q
    qwen3-vl-235b-a22b-instruct
    ToolsJSONVision17.0 intel · $0.200/M · 36 t/s
    17.0
    Intelligence
  23. 23Q
    qwen3-vl-8b-thinking
    ReasoningToolsJSON+116.7 intel · $0.117/M · 100 t/s
    16.7
    Intelligence
  24. 24Q
    qwen3-32b
    ReasoningToolsJSON16.5 intel · $0.080/M · 303 t/s
    16.5
    Intelligence
  25. 25Q
    qwen3-14b
    ReasoningToolsJSON16.2 intel · $0.100/M · 59 t/s
    16.2
    Intelligence

Frequently asked

What is the best open-source Qwen model?

The best open-weight Qwen model is Qwen3.5 397B A17B (40.1 intelligence) — downloadable from Hugging Face and self-hostable. Qwen3.5-27B (37.2) and Qwen3.6 27B (37.1) round out the top three.

What's a good alternative to Qwen3.5 397B A17B?

Qwen3.5-27B (37.2) is the closest alternative on this metric, followed by Qwen3.6 27B (37.1). See the full ranking above for the tradeoffs.

How many Qwen models are there?

modelgrep tracks 49 Qwen models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by Qwen3.7 Max. 25 of them qualify for this ranking.

More Qwen rankings

All rankings