DeepSeek V3.1on Krater.ai
DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context
Released August 21, 2025
Specifications
DeepSeek V3.1 at a glance
Context
163.8K tokens
Max output
147.5K tokens
Input
Text
Output
Text
Input price
$0.25 / 1M
Output price
$0.95 / 1M
Knowledge cutoff
2025-03-31
Is DeepSeek down?
DeepSeek timing out? The live DeepSeek status page shows current incidents and faster models to switch to.
Live DeepSeek statusCapabilities
What it is good at
Catalog capabilities
- Generates text responses
- Supports tool calling
- Supports structured outputs
- Instruction format: deepseek-v3.1
Good for
Benchmarks
Catalog data, not Krater benchmarks.
Design Arena svg rank
#97
Design Arena dataviz rank
#105
Compare
Compare DeepSeek V3.1
DeepSeek V3.1 vs Chat
DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions.
Open comparisonDeepSeek V3.1 vs DeepSeek V3 0324
V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team.
Open comparisonDeepSeek V3.1 vs DeepSeek R1
Performance on par with OpenAI o1, but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active in an inference pass
Open comparisonFAQ
Questions about DeepSeek V3.1
Clear answers based on the model catalog and Krater access.
Use DeepSeek V3.1 in your API
Connect with Krater's unified API.