DeepSeek launches V4.1-Flash with native multimodal, tests show it beats V4-Pro
teortaxesTex · x · 2026-10-22
A DeepSeek API announcement (screenshot) says V4.1-Flash is live with native multimodal support — set the model to deepseek-flash.
Key points:
- V4-Flash and V4-Flash-Vision-Exp are retired; deepseek-v4-flash and deepseek-v4-flash-vision-exp temporarily route to V4.1-Flash for compatibility.
- Tests by multiple parties put V4.1-Flash ahead of V4-Pro on performance, cost, speed and total runtime; V4-Pro is being phased out.
- From 04:00 UTC on Sept 14, 2026, all deepseek-v4-pro requests will route to V4.1-Flash at V4.1-Flash rates until V4.1-Pro launches.
- Official partners WorkBuddy (including CodeBuddy) and OpenCode fully support V4.1-Flash.
Poster teortaxesTex questions how the OpenCode partnership economics work: given DeepSeek's margins are grounded in GPU physical limits, is OpenCode hosting independently, or is it a side channel to sell DeepSeek subscriptions without DeepSeek breaking its 'never sell subscriptions' stance?
Related event: DeepSeek V4.1-Flash Tops Open-Source Rankings in Agent and WebDev Arenas(8 posts)→
More from Models
- OpenAI Voice Mode Users Hit Long Wait Queues Even on Pro Plan — deviltalk · 2026-10-22
- LLMs predict MtG set win rates pre-release; Claude Opus 5.5 does best — Jsevillamol · 2026-10-22
- Claude models feel 'so far ahead right now,' says popular AI commentator — iruletheworldmo · 2026-10-22
- Rumor: Anthropic's 'Claude Duet' is a proactive personal assistant codenamed Dot — LeviTurk · 2026-10-22
- DeepSeek's V4.1 pricing may starve rivals and squeeze neoclouds, analyst argues — teortaxesTex · 2026-10-22
- php2shell was found by GPT-5.6 Sol, not Mythos, sparking open-source frontier model risk debate — xeophon · 2026-10-22