LLM Reference

PixArt-Σ

Released
2024-03-30
Last refreshed
2026-05-19
Status
Researched 38d ago
Open sourceCommercial use: permittedMultimodalVision

PixArt-Σ is a released vision model with open-source; evaluate it while provider pricing coverage matures.

Use it for

  • Teams evaluating vision

Do not use it for

  • Cost-sensitive launches that need sourced token pricing
  • Strict JSON or tool-calling flows
  • Teams that need a tracked hosted API route today
Specifications
Family
PixArt
Released
2024-03-30
Parameters
600M
Architecture
Diffusion Transformer
Specialization
image-generation
Openness
Open source
License
Apache 2.0OSI-approvedCommercial use: permitted
Created by

Bridging technology and business through AI

Hong Kong, China
Founded 2012
Website
Pricing

No tracked provider token pricing is available yet.

About

PixArt-Σ (Sigma) by Huawei Noah's Ark Lab. A 600M-parameter Diffusion Transformer capable of generating images at up to 4K (3840×2560) resolution without an upscaler. Uses 'weak-to-strong training' on higher-quality data. Competitive with Adobe Firefly 2, Imagen 2, and DALL-E 3 at a fraction of the parameter count. Open source under Apache 2.0. Available via Fal API (fal-ai/pixart-sigma).

PixArt-Σ is an open-source model in the PixArt family. The structured metadata tracks multimodal input. No headline benchmark score is tracked for PixArt-Σ yet.

Top use-case fit

Vision

Included by capability and metadata signals in the decision map.

Provider price ladder

No tracked provider token pricing is available for this model yet.

Capabilities

Multimodal

Benchmark peer barsfor Vision

No task-mapped benchmark peers are available for this model yet.

Migration checks

No linked migration route is available for this model yet.