DeepSeek: DeepSeek V3.1

DeepSeek

DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinki

High value
Text
Visit official site →Compare →

Specifications

VendorDeepSeek
Familydeepseek
Context163.8K · 122.9k words（≈ a thesis）
Max output163.8K · 122.9k words（≈ a thesis）
Released2026-06-19
Open sourceNo
Tool use—

Pricing across channels

per 1M tokens, USD

ChannelTypeInputOutput→
OpenRouter (auto)Aggregator$0.21$0.79visitAd
OpenRouter API (auto-fetched)· 2026-06-19

Related models

Similar tier, vendor or price range

DeepSeek: DeepSeek V3
DeepSeek
Intel. —$0.20/$0.80
DeepSeek: DeepSeek V3 0324
DeepSeek
Intel. —$0.20/$0.77
DeepSeek: DeepSeek V3.1 Terminus
DeepSeek
Intel. —$0.27/$0.95
DeepSeek: DeepSeek V3.2 Exp
DeepSeek
Intel. —$0.27/$0.41