NewKimi K3 is now live, hosted on PrivateMind infrastructure.Try it free
Sign up

Models

Leading open-weight models, served from hardware we control

Open-weight models run on PrivateMind infrastructure, with nothing routed to a public cloud and nothing you send used for training. Frontier models from Claude, OpenAI and Amazon Bedrock are available through hybrid routing, which is off until your firm turns it on.

Open the playgroundBrowse the catalogue
  • Moonshot AI
  • Z.ai
  • Qwen
  • Mistral
  • NVIDIA
  • Resemble AI
  • Ai2
Plus hybrid routing toClaudeOpenAIAmazon Bedrock

Try the models

Send up to three prompts free, no account needed. Then sign up for an API key.

Image, speech, OCR and embedding models are available through the API. Full catalogue

Kimi K3 2.8T-104B

What are we figuring out?

Own your data. Own your intelligence.

Prefer to start in code? Read the API quickstart or the Python SDK docs.

Catalogue

Full catalogue

Speech, documents, images and retrieval run on the same infrastructure and the same API key. The catalogue can differ between organisations, so call GET /v1/models to see what is live for you. Aliases such as fast and reasoning are explained on the Platform page.

Model reference
Chat and vision

Qwen 3.8 27B

The everyday default: fast dense 27B with vision, thinking and tools

  • Vision
  • Tools
  • Reasoning
  • Open weights
Model ID
qwen3-8-27b-nvfp4
Context
262K
Per million tokens
On request
Try in playgroundSign up to deploy
Chat

GLM 5.3 753B-40B

Frontier coding and agentic work, NVFP4 for Blackwell

  • Tools
  • Reasoning
  • Open weights
Model ID
glm-5-3-nvfp4
Context
524K
Per million tokens
On request
API docsSign up to deploy
Chat and vision

Kimi K3 2.8T-104B

Frontier agentic coding, native vision, 500K context

  • Vision
  • Tools
  • Reasoning
  • Open weights
Model ID
kimi-k3
Context
500K
Per million tokens
On request
Try in playgroundSign up to deploy
Chat and vision

GLM 5.3 Flash 320B-18B

Fast agentic coding with vision, 320B MoE in FP8

  • Vision
  • Tools
  • Reasoning
  • Open weights
Model ID
glm-5-3-flash-fp8
Context
524K
Per million tokens
On request
Try in playgroundSign up to deploy
Hybrid routing

Frontier models

Claude, OpenAI and Amazon Bedrock models, routed through the PrivateMind chat app and API with the same keys and controls.

Availability
Off until your organisation enables it
Per million tokens
On request
Ask about hybrid routing
Reranking

Qwen3 Reranker 4B

Re-ranks retrieval hits by relevance

  • Open weights
Model ID
qwen3-reranker-4b
Context
8K
Per million tokens
On request
API docsSign up to deploy
Embeddings

Nemotron 3 Embed 8B

Multilingual text embeddings for retrieval

  • Open weights
Model ID
nemotron-3-embed-8b
Context
33K
Per million tokens
On request
API docsSign up to deploy
Speech to text

Voxtral Realtime 4B

Real-time multilingual speech-to-text

  • Open weights
Model ID
voxtral-mini-4b-realtime
Context
33K
Per million tokens
On request
API docsSign up to deploy
Text to speech

Chatterbox Turbo 350M

Low-latency streaming text-to-speech

  • Open weights
Model ID
chatterbox-turbo
Per million tokens
On request
API docsSign up to deploy
Image generation

Cosmos 3 Super 64B

Generates images from text prompts

  • Open weights
Model ID
cosmos3-super-text2image
Per million tokens
On request
API docsSign up to deploy
Safety

Qwen3Guard 8B

Flags unsafe prompts and responses

  • Open weights
Model ID
qwen3guard-gen-8b
Context
33K
Per million tokens
On request
API docsSign up to deploy
Document OCR

olmOCR 2 7B

Document OCR to markdown with HTML tables and LaTeX

  • Open weights
Model ID
olmocr-2-7b-1025-fp8
Context
16K
Per million tokens
On request
API docsSign up to deploy

Pricing

Model pricing

Per-model rates on pooled capacity are available on request. Batch requests are half price. See pricing for credits and enterprise options.

Need a model that is not in the catalogue?

With dedicated capacity you can deploy catalogue models or your own weights as private, OpenAI-compatible endpoints.

Start free today, or request a demo of private AI for your firm

Create an account to use chat and the API straight away, or request a demo to see dedicated capacity, workspaces and engineers for your team.

Sign upRequest a demo