DeepSeek launches V4.1-Flash with encoder-decoder architecture and native vision
ricklamers · x · 2026-09-10
DeepSeek officially introduced DeepSeek-V4.1-Flash, the smallest model in its new architecture family, touting native visual understanding, faster inference, and higher throughput with an eye toward scaling up. A quoted thread by inference developer Halex623 lists key architecture changes: encoder-decoder design, an "engram" memory module, different active parameter counts for prefill vs. decode, a newer mHC, and a new sparse indexer — noting the list isn't exhaustive and that supporting it in inference will be a challenge.
More from Infra
- Michael Dell: Musk is "the most demanding customer there is" — Colossus servers built in ~6 weeks — elonmusk · 2026-10-27
- AI chip startup Tenstorrent raises $1.5B at $6.2B valuation, more than doubling — Beth_Kindig · 2026-10-27
- Anthropic discloses up to $120.5B SpaceX compute deal, ~$3.1B monthly at full capacity — XFreeze · 2026-10-27
- Satire: swapping Raft for an LLM-style decision model tanks write throughput 99.98% — irombie · 2026-10-27
- Cadence joins the domain-specific AI model camp for chip design — karlfreund · 2026-10-27
- Qdrant builds Universal I/O to run vector search mmap code on object storage — qdrant_engine · 2026-10-27