Open-source MOSS-Image tops 1M ComfyUI downloads, 5B model rivals closed rivals
songguoxiansen · x · 2026-10-23
The 5B-parameter open-source text-to-image model MOSS-Image has surpassed 1 million downloads on ComfyUI. A hands-on test with 10 popular prompts covered scenarios from portrait photography and concert screens to infographics and character turnaround sheets, with the full ComfyUI workflow (Qwen3.5-9B as CLIP, FLUX.2 VAE) shown in the attached image.
The verdict: for a 5B open-source model, output quality is impressively close to some closed models. A standout feature is structured JSON Layout control, letting users specify exact coordinates for text and visual elements — Apache 2.0 licensed.
More from Multimodal
- FLUX.3 Image Klein tested across 15 complex prompts: strong on text, weak on distant faces — SkirtSwimming8950 · 2026-10-23
- Two Midjourney sref codes for grainy pastel fashion illustration style — sergeantsref · 2026-10-23
- Three free ways to try MOSS-Image: API playground, HF Space, ComfyUI — songguoxiansen · 2026-10-23
- Testing light effects in MiniMax video generation — the guinea pig was fine — seppe0815 · 2026-10-23
- GAFU Uses Generation-Side Representations to Boost Fine-Grained Visual Understanding in Unified Multimodal Models — Tengju Ru · 2026-10-23
- MiniMax Design MCP goes live: create with Design inside Codex and other agents — Hailuo_AI · 2026-10-23