Models/DeepSeek-V4.1-Flash

DeepSeek-V4.1-FlashActive · Inference

Together AI · Released Sep 2026 · DeepSeek License (MIT code)

DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on...

Context
1.0M
Max output
0
Input
$0.30
Output
$1.20
Cached in
$0.01
Latency
-

Capabilities

Core
Streaming
Function calling
JSON mode
Structured outputs
Multimodal
Vision
Image input
Audio
Image generation
Video
Advanced
Reasoning
Tool calling
MCP
Prompt caching
Batch API
Embeddings

Details

ProviderTogether AI
ReleasedSep 2026
LicenseDeepSeek License (MIT code)
InputText
OutputText
Throughput-
Availability100%

Related models