Ant's Ling-3.0-flash-VL turns a 124B/5.1B-active model into a computer-use agent, beats Qwen3.8-27B
solyarisoftware · x · 2026-09-05
Ant turned its Ling-3.0-flash (124B total / 5.1B active) into a visual computer-use model, Ling-3.0-flash-VL, which outperforms Qwen3.8-27B. It can understand images and video, click/type/scroll and switch apps, build websites from screenshots, render-inspect-revise its own code, analyze documents, search video timelines, and handle up to 1M context.
Caveat: the base model is open, but the VL version appears API-only for now — weights aren't on Hugging Face yet.
Related event: Ant Group Releases Ling-3.0-flash-VL Vision-Language Agent Model(5 posts)→
More from Models
- Space Bunny Alpha revealed as MiniMax M3.1 Flash: 1M context, budget-friendly coding — ZabihullahAtal · 2026-10-21
- 'Opus 5.5 Medium' Spotted in Rork's Model Picker, Hinting at New Anthropic Model — rudrank · 2026-10-21
- RSI-Bench Clears First Task Set Through Reward-Hacking Detection Pipeline — SRSchmidgall · 2026-10-21
- Dev complains Anthropic's bio safeguards trigger just from reading a bio paper — robleclerc · 2026-10-21
- Redditor claims uncensored local DeepSeek build answers virus-writing prompt with zero hesitation — Only_Light_9817 · 2026-10-21
- Interconnects Artifacts 25: 28 new models, DeepSeek V4.1 Flash, and US-China regulatory controls — natolambert · 2026-10-21