Benchmark Almanac

Qwen

Best on the Capabilities Index Qwen 3.8 Max 158.0
94 models 32 model cards read qwen.ai

Qwen's models against the field

Every model on the Epoch Capabilities Index by release date, with Qwen's in color.

50 75 100 125 150 175 2023 2024 2025 2026 2027 Qwen3.8 Max (0902) Qwen3.7-Plus 50 75 100 125 150 175 2023 2024 2025 2026 2027 Qwen3.8 Max (0902) Qwen3.7-Plus

Its strongest models

Qwen Aug 2026

Qwen 3.8 Max

Closed weights Flagship

158.0 #22 of 271
Epoch Capabilities Index
  • IFBench 82.8% #2 /70
  • MMMU-Pro 82.3% #4 /72
  • DeepSWE 57.5% #16 /28
Qwen Sep 2026

Qwen3.8 Max (0902)

Closed weights

156.2 #32 of 271
Epoch Capabilities Index
  • Terminal-Bench 4.0 27% #13 /23
  • CritPt 17.7% #39 /160
  • Vals Finance Agent v2 57.9% #12 /80
Qwen May 2026

Qwen3.7-Max

Closed weights Flagship

154.2 #45 of 271
Epoch Capabilities Index
  • SWE-bench Verified 77.3% #7 /32
  • IFBench 79.1% #12 /70
  • SciCode 53.5% #6 /47
Qwen Aug 2026

Qwen 3.8 27B

Open weights

151.8 #55 of 271
Epoch Capabilities Index
  • IFBench 79.5% #11 /70
  • NL2Repo-Bench 42.3% #19 /28
  • ERQA 65.5% #12 /26
Qwen Apr 2026

Qwen 3.6 Max (Preview)

149.4 #66 of 271
Epoch Capabilities Index
  • SWE-bench Verified 76.7% #11 /32
  • SciCode 47% #17 /47
  • NL2Repo-Bench 42.9% #17 /28
Qwen Mar 2026

Qwen 3.6 Plus

Closed weights Flagship

147.8 #72 of 271
Epoch Capabilities Index
  • SWE-bench Verified 57.9% #29 /32
  • IFBench 74.2% #25 /70
  • MMMU-Pro 78.8% #17 /72

What its model cards measure

Tall ticks are Qwen model cards that list the benchmark in their headline table. Gray dots are cards that leave it out. Rows are the benchmarks the lab listed most, at least twice each, in the order it started using them.

C-Eval GPQA Diamond MMLU-Pro IFEval MMLU-Redux LiveBench MATH-Vision (MathVision) MMMU-Pro Video-MME LVBench MLVU MMMU MMStar MathVista AI2D AndroidWorld RealWorldQA SWE-bench Verified SWE-bench Multilingual Humanity's Last Exam LiveCodeBench v6 SuperGPQA HMMT 2025 CharXiv ERQA IFBench SWE-Bench Pro NL2Repo-Bench

Model cards

DateModelDocumentBenchmarks
18 Sep 2026 Qwen3.8-Omni-Flash Qwen3.8-Omni-Flash: Omni Senses. Agentic Delivery. (blog) 45
26 Aug 2026 Qwen3.8 Flash Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-Efficiency (blog) 22
14 Aug 2026 Qwen 3.8 27B Qwen3.8-27B model card (Hugging Face) 25
3 Aug 2026 Qwen 3.8 Max Qwen3.8-Max: A New Bar for Coding and Cowork (blog) 84
22 Jun 2026 Qwen-AgentWorld-35B-A3B Qwen-AgentWorld-35B-A3B model card (Hugging Face) 1
1 Jun 2026 Qwen3.7-Plus Qwen3.7-Plus: Multimodal Agent Intelligence (blog) 73
20 May 2026 Qwen3.7-Max Qwen3.7: The Agent Frontier (blog) 40
22 Apr 2026 Qwen3.6 27B Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model (blog) 45
20 Apr 2026 Qwen 3.6 Max (Preview) Qwen3.6-Max-Preview: Smarter, Sharper, Still Evolving (blog) 12
15 Apr 2026 Qwen 3.6 35B-A3B Qwen3.6-35B-A3B: Agentic Coding Power, Now Open to All (blog) 50
2 Apr 2026 Qwen 3.6 Plus Qwen3.6-Plus: Towards Real World Agents (blog) 64
30 Mar 2026 Qwen3.5-Omni-Plus Qwen3.5-Omni: Scaling Up, Toward Native Omni-Modal AGI (blog) 63
24 Feb 2026 Qwen3.5-122B-A10B Qwen3.5-122B-A10B model card (Hugging Face), Qwen3.5 medium series 78
16 Feb 2026 Qwen3.5 397B-A17B Qwen3.5-397B-A17B model card 79
2 Feb 2026 qwen3-coder-next Qwen3-Coder-Next: Pushing Small Hybrid Models on Agentic Coding (blog) 5
23 Jan 2026 Qwen3-Max Pushing Qwen3-Max-Thinking Beyond its Limits (blog) 18
24 Sep 2025 Qwen3-Max Qwen3-Max: Just Scale it (blog) 8
23 Sep 2025 Qwen3-VL-235B-A22B Qwen3-VL launch post (qwen.ai blog) 64
22 Sep 2025 Qwen3-Omni-30B-A3B Qwen3-Omni-30B-A3B-Instruct model card (Hugging Face) 37
11 Sep 2025 Qwen3-Next-80B-A3B Qwen3-Next-80B-A3B-Thinking model card (Hugging Face) 20
25 Jul 2025 Qwen3-235B-A22B-Thinking Jul 2025 Qwen3-235B-A22B-Thinking-2507 model card 22
22 Jul 2025 Qwen3-Coder-480B-A35B Qwen3-Coder: Agentic Coding in the World (blog) 12
29 Apr 2025 Qwen3-235B-A22B Qwen3 blog (image) and Qwen3 Technical Report, Table 11 24
27 Mar 2025 Qwen2.5-Omni-7B Qwen2.5-Omni-7B model card (Hugging Face) 42
6 Mar 2025 QwQ-32B QwQ-32B: Embracing the Power of Reinforcement Learning (blog) and HF model card 5
28 Jan 2025 Qwen2.5-VL-72B Qwen2.5-VL-72B-Instruct model card (Hugging Face) 29
28 Jan 2025 Qwen2.5-Max Qwen2.5-Max: Exploring the Intelligence of Large-scale MoE Model (blog) 14
27 Jan 2025 Qwen2.5-14B-Instruct-1M Qwen2.5-1M: Deploy Your Own Qwen with Context Length up to 1M Tokens (blog) 4
19 Sep 2024 Qwen2.5-72B Qwen2.5-LLM: Extending the boundary of LLMs (blog) 34
7 Jun 2024 Qwen2-72B Hello Qwen2 (blog) 28
5 Feb 2024 Qwen1.5-72B Introducing Qwen1.5 (blog) 19
25 Sep 2023 Qwen-14B Qwen-14B and Qwen-14B-Chat model cards 12

Every Qwen model