DeepSeek-V4.1-Flash
The first and smallest model in DeepSeek's new architecture family — natively multimodal, and now what the deepseek-flash API name serves
DeepSeek-V4.1-Flash
DeepSeek • September 2026
Training Data
Not disclosed
DeepSeek-V4.1-Flash
September 2026
Parameters
Not disclosed
Training Method
New architecture family built for a higher capability ceiling, faster inference, and higher throughput
Context Window
1,000,000 tokens
Knowledge Cutoff
Not disclosed
Key Features
1M Context / 384K Output • Native Vision • Peak/Off-Peak Pricing • Anthropic & Responses APIs
Capabilities
Coding: Outstanding (90.6% Terminal-Bench 2.1)
Reasoning: Outstanding (90.9 GPQA Diamond)
Cost Efficiency: Maximum
What's New in This Version
Adds native image understanding and posts 74.2 on DeepSWE v1.1 and a 3471 Codeforces rating, while pricing below V4-Flash: $0.15 in / $0.60 out per MTok off-peak, $0.30 / $1.20 at peak
The first and smallest model in DeepSeek's new architecture family — natively multimodal, and now what the deepseek-flash API name serves
What's New in This Version
Adds native image understanding and posts 74.2 on DeepSWE v1.1 and a 3471 Codeforces rating, while pricing below V4-Flash: $0.15 in / $0.60 out per MTok off-peak, $0.30 / $1.20 at peak
Technical Specifications
Key Features
Capabilities
Other DeepSeek Models
Explore more models from DeepSeek
DeepSeek-V4-Pro
DeepSeek's frontier MoE flagship, generally available since August 2026 with substantially stronger agentic tool use and code execution
DeepSeek-V4-Flash
DeepSeek's smaller, fast variant of V4 — same architecture at a fraction of the cost and latency
DeepSeek-V3.2
DeepSeek's latest flagship model matching GPT-5 performance with integrated tool-use thinking
