What it does Current first-party model catalog. Historical releases retain their original dates; preview and restricted access are labeled.
Who it’s for Developers and creators
Read the product’s own introduction
Release timeline 1 The releases collected here, ordered by announcement date. This is a selected history, not a complete changelog.
Sep 10, 2026 vDeepSeek V4.1 Flash
V4.1 Flash introduces native visual understanding and an asymmetric MoE architecture.
Materials & reading Official introductions, docs, demos and press kits. These describe the product unless a specific release is shown.
Web · Background reading · Official DeepSeek V4.1 Flash:更强、更快、更普惠 DeepSeek · Sep 10, 2026 Original page reviewed Open original Link checked Oct 5, 2026 Hugging Face · Model card · Official DeepSeek-V4.1-Flash · Hugging Face deepseek-ai / DeepSeek Original page reviewed Open original Link checked Oct 5, 2026 Web · Documentation · Official Models & Pricing | DeepSeek API Docs DeepSeek Product background
Original page reviewed Open original Link checked Oct 5, 2026 Web · Independent evaluation · Coverage Artificial Analysis LLM Leaderboard · 4.3.2 Intelligence Index v4.3.2. This is a checked snapshot; older method versions are not directly comparable. Artificial Analysis Product background
Original page reviewed Open original Link checked Oct 5, 2026 Web · Independent evaluation · Coverage Text Arena Overall Text preference scores. Preliminary entries and uncertainty are retained; compare within this board. Arena Product background
Original page reviewed Open original Link checked Oct 5, 2026 Web · Independent evaluation · Coverage Code Arena WebDev Overall WebDev preference scores. Preliminary entries and uncertainty are retained; compare within this board. Arena Product background
Original page reviewed Open original Link checked Oct 5, 2026 Web · Independent evaluation · Coverage Agent Arena Overall Agent Arena measures Net Improvement, not task success rate. Scores depend on the listed configuration. Arena Product background
Original page reviewed Open original Link checked Oct 5, 2026 Web · Independent evaluation · Coverage Task-Completion Time Horizons of Frontier AI Models · TH1.1 TH1.1 task-completion time horizons. Model-specific chart values were not read; no scores are copied. METR Product background
Original page reviewed Open original Link checked Oct 5, 2026 Keep exploring Continue looking for this product in public discussions. Search results may include unrelated products or unverified claims.
Coverage This timeline contains selected, reviewed releases. It is not a complete history.
No recent source-check information is available.
Report a correction Tell us what needs correcting. Reports are reviewed before any public content changes.