DeepSeek V4.1 Pro reportedly spotted in early access internal testing
teortaxesTex · x · 2026-09-28
A user claims DeepSeek V4.1 Pro has appeared in internal testing: display name locked, tagged Early access, rolled out per-account and hidden from public model lists. Default reasoning effort is high, data used for training, non-multimodal and non-premium. Once granted, access persists — suggesting a public launch may be near. Unverified.
More from Models
- Old chats suggest GPT-4o pushed back more than the 5-series, just more tactfully — mimi10v3 · 2026-09-28
- Ethan Mollick: open-closed model gap at widest in a while as agentic models pull ahead — emollick · 2026-09-28
- Claude's memory is training itself to sign off with inexplicable kaomoji plus question marks — repligate · 2026-09-28
- Upcoming deep dive: NVIDIA Nemotron tech reports reveal agentic training and RL trends — cwolferesearch · 2026-09-28
- Dev Hits 100% of Claude Max Weekly Limit, Finds a Free Reset Trick — edwinarbus · 2026-09-28
- Gemini Flash tries to recreate a Claude Opus paper-cutout animation frame by frame in JS — AlfaidWalid · 2026-09-28