DeepSeek-V4.1-Flash: Smarter, Faster, More Efficient
Mainly relevant if your organisation builds on DeepSeek's API — it's a cheaper, more efficient model swap that should lower your running costs, but note existing V4-Pro API calls will start being billed at the new V4.1-Flash rate from 14 September 2026, so check your usage before then.
What changed
Source summary
DeepSeek released V4.1-Flash, a new MoE model (552B total parameters, 8B/16B active) with native multimodal support and much lower KV-cache memory and storage needs, priced lower than its predecessor. Existing V4-Pro API traffic will be automatically routed to V4.1-Flash pricing from 14 September 2026.
Why it matters to your work
Practical interpretation
Mainly relevant if your organisation builds on DeepSeek's API — it's a cheaper, more efficient model swap that should lower your running costs, but note existing V4-Pro API calls will start being billed at the new V4.1-Flash rate from 14 September 2026, so check your usage before then.
Source & availability
This is a summary of a published provider announcement, with practical context from SCALA’s feed.
Plan, region, workspace and rollout conditions may affect access in Australia. Check the original source for current availability.
api-docs.deepseek.comThis update is primarily for technical teams.