LLM Reference

DeepSeek V4.1 Models by DeepSeek

DeepSeekMITOpen source
1 model2026Up to 1m ctxFrom $0.15/1M input

Last refreshed 2026-09-10. Next refresh: weekly.

Details

ResearcherDeepSeek
LicenseMITOSI-approved
Commercial useCommercial use: permitted
Models1
Released2026
Max context1m

Capabilities

VisionAll models
MultimodalAll models
ReasoningAll models
JSON / Tool useAll models
Structured OutputsAll models

About

DeepSeek V4.1 is the September 2026 DeepSeek release built on a 552B-parameter mixture-of-experts architecture with Compressed Expert Dispatch (CED). The family launches with DeepSeek V4.1 Flash, an open-weights multimodal model with 1M-token context, vision input, reasoning, and tool use. Direct DeepSeek API off-peak pricing starts at $0.15 / $0.60 per million input / output tokens.

Decision facts

Best fit
vision and multimodal workreasoningJSON / Tool use
Capability starting point
DeepSeek V4.1 Flash with 1m context and reasoning, JSON / Tool use, structured outputs, and multimodal inputs
Lowest tracked input
DeepSeek V4.1 Flash · $0.15/1M · DeepSeek Platform
Closest related family
Janus

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

1 in view

Use when the workload needs 1m context, 552B parameters, and reasoning.

2026-091m context552B parametersreasoning

Release Timeline

1 release group
2026-09
1 current
DeepSeek V4.1 Flash
1m context552B parametersreasoning
Current

Specifications(1 models)

DeepSeek V4.1 model specifications comparison
ModelReleasedContextParametersVisionMultimodalReasoningJSON / Tool useStructured Outputs
DeepSeek V4.1 Flash2026-091m552BYesYesYesYesYes

Available From(2 providers)

Pricing

DeepSeek V4.1 model pricing by provider
ModelProviderInput / 1MOutput / 1MType
DeepSeek V4.1 FlashDeepSeek Platform$0.15$0.6Serverless
DeepSeek V4.1 FlashOpenRouter$0.15$0.6Serverless