AI Model Timeline

Tracking the accelerating release frequency of frontier AI models.

Qwen-Image-3.0

Alibaba·

Imageimagecloud

Overview

Third-generation text-to-image model from Alibaba's Qwen team, focused on dense text and layout rendering. Accepts prompts up to 4,500 tokens, renders in-image text in 12 languages and 20+ fonts down to ~10px, and produces complex multi-element layouts (newspapers, storyboards, infographics, UI mockups, knowledge graphs) in a single pass, generating up to 9 images at once. Unlike the earlier Apache 2.0 releases (Qwen-Image 1.0, Aug 2025, and 2.0, each with a technical report), 3.0 shipped without open weights, benchmarks or a technical report. Available via chat.qwen.ai, Qwen Studio and Alibaba's API platform; API pricing not disclosed at launch.

Capabilities and innovations

Text-to-ImagePrompts up to 4,500 tokensIn-image text rendering (12 languages, 20+ fonts, ~10px)Single-pass complex layoutsUp to 9 images per generationLong-prompt layout generation (4.5K tokens)High-density multilingual text renderingClosed release (no open weights, unlike prior Qwen-Image versions)

Benchmarks

No benchmark values are recorded for this model.

  • GPQA Diamond: not reported by the vendor
  • Humanity’s Last Exam (no tools): not reported by the vendor
  • SWE-bench Verified: not reported by the vendor
  • SWE-bench Pro: not reported by the vendor
  • MMLU: not reported by the vendor
  • MMLU-Pro: not reported by the vendor
  • Terminal-Bench 2.x: not reported by the vendor
  • MMMU-Pro: not reported by the vendor

Links

More from Alibaba

Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.