All models
ReasoningComing soon
DeepSeek 4.1 Flash
Think bigger. Go deeper.
MIT-licensed, with Terminal-Bench 2.1 results above leading frontier models and reasoning across a million tokens.
Good for
- 01Analyze large codebases and complex technical documentation.
- 02Perform sophisticated reasoning across up to 1M tokens.
- 03Build advanced analytical workflows and AI agents.
- Coding
- Data & analytics
- Documentation
- Information systems & search
Premium reasoning when your workloads demand more. Scale with transparent token-based pricing.
Key facts
- Model origin
- DeepSeek
- Context window
- 1M tokens
- Size
- 552B MoE · 8B / 16B active
- Modalities
- text
- API model id
- Model id published at launch
- Availability
- Coming soon. Access opens through the waitlist.
Your data boundary
Prompts and outputs are not stored by default. Knowledge or Memory data that your organization intentionally persists is encrypted, pinned to the selected region, isolated to authorized organizational scopes, and never used to train shared or third-party models.
Works well with
EmbeddingsAvailable nowQwen3 Embedding 0.6BMultilingual search across 100+ languagesReasoningComing soonGLM 5.3Advanced reasoning & agentic codingChatAvailable nowQwen 3.8Enterprise chat, coding & 1M context
More models
Qwen 3.8Gemma 4ClefFLUX.2 [klein]Whisper large v3mxbai-embed-large v1Qwen3 Embedding 0.6BKokoroGLM 5.3
Included unlimited* on every plan*
Explore the platform
Start with control
Create your account and pick the region your data must stay in.