Skip to main content

Model Comparison

Schatzi AI provides access to 20 active state-of-the-art AI models. All models run exclusively on Swiss infrastructure, ensuring your data never leaves Switzerland and remains compliant with Swiss data protection regulations.

Complete Model Reference​

Swiss LLM (AI Act Compliant)​

Apertus Swiss LLM - Large​

  • Context Window: 65,536 tokens
  • Streaming: Supported
  • Availability: Chat UI & API

Ideal for multilingual services, government agencies and R&D teams looking for a reliable, adaptable model ● Data and methods documented for unprecedented transparency ● Compliant with the AI Act and respectful of privacy and intellectual property ● A 70B version with performance on a par with current market leaders

Capabilities: Chat, Multi-lingual, Swiss LLM

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Apertus Swiss LLM - Large (v1.5)​

  • Context Window: 65,536 tokens
  • Streaming: Supported
  • Availability: API only

Apertus v1.5. Ideal for multilingual services, government agencies and R&D teams looking for a reliable, adaptable model ● Data and methods documented for unprecedented transparency ● Compliant with the AI Act and respectful of privacy and intellectual property ● A 70B version with performance on a par with current market leaders

Capabilities: Chat, Multi-lingual, Swiss LLM

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Reasoning & Problem-Solving​

Chat & Document Analysis & Reasoning - Large​

  • Context Window: Not specified
  • Streaming: Supported
  • Availability: API only

large-scale model, rivalling leading models or leading models Opus across a broad range of complex tasks ● Advanced multilingual capabilities ● Reasoning mode can be enabled to dynamically tailor responses to the context and complexity of queries

Capabilities: Document Analysis, Chat, Vision, Reasoning, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Fast Reasoning & Instruction Following - Small​

  • Context Window: 32,768 tokens
  • Streaming: Supported
  • Availability: API only

Optimized for Reasoning and instruction-following capabilities

Capabilities: Thinking, Chat, Data Analysis, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Reasoning & Problem Solving - Small​

  • Context Window: 32,768 tokens
  • Streaming: Supported
  • Availability: API only

Optimized for thinking and reasoning

Capabilities: Thinking, Chat, Reasoning, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Reasoning & Agent tasks - Large​

  • Context Window: 65,536 tokens
  • Streaming: Supported
  • Availability: API only

Optimized for powerful reasoning, agentic tasks, and versatile developer use cases

Capabilities: Data Analysis, Chat, Thinking, Agent, Reasoning, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Reasoning & Problem Solving - Medium​

  • Context Window: 32,768 tokens
  • Streaming: Supported
  • Availability: API only

Optimized for thinking and reasoning

Capabilities: Thinking, Chat, Reasoning, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Reasoning & Problem Solving - Xtra Large​

  • Context Window: Not specified
  • Streaming: Supported
  • Availability: API only

Optimized for Reasoning chat completions. Reasoning model

Capabilities: Thinking, Chat

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Reasoning & Tool Use - Large (GLM-4.5 Air)​

  • Context Window: 131,072 tokens
  • Streaming: Supported
  • Availability: API only

ZhipuAI GLM-4.5-Air. Mixture-of-Experts model with 106B total / 12B active parameters. Hybrid reasoning with configurable thinking mode, strong tool/function calling and code generation capabilities. 128K context window.

Capabilities: Thinking, Chat, Function Calling, Reasoning

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Chat, Document Analysis & Agent tasks - Xtra Large​

  • Context Window: 250,000 tokens
  • Streaming: Supported
  • Availability: Chat UI & API

Very large-scale model, rivalling leading models or leading models Opus across a broad range of complex tasks ● Advanced multilingual capabilities ● Reasoning mode can be enabled to dynamically tailor responses to the context and complexity of queries ● Optimized for powerful reasoning, agentic tasks, and versatile developer use cases

Capabilities: Chat, Document Analysis, Agent, Coding, Thinking, Web Search, Vision, Reasoning, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Chat, Document Analysis, Coding & Reasoning - Xtra Large​

  • Context Window: 1,000,000 tokens
  • Streaming: Supported
  • Availability: Chat UI & API

Multi modal model, optimized for chat, document analysis, coding and reasoning.

Capabilities: Chat, Document Analysis, Coding, Thinking, Data Analysis, Vision, Reasoning, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Chat, Vision, Document Analysis & Reasoning - Medium​

  • Context Window: 256,000 tokens
  • Streaming: Supported
  • Availability: Chat UI & API

Best in class multi-modal model, optimized for chat, vision, document analysis, coding and reasoning.

Capabilities: Chat, Vision, Document Analysis, Coding, Thinking, Reasoning, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Reasoning & Tool Use - Xtra Large (GLM-5.2)​

  • Context Window: Not specified
  • Streaming: Supported
  • Availability: Chat UI & API

ZhipuAI GLM-5.2 Hybrid reasoning with configurable thinking mode, strong tool/function calling and code generation capabilities. Same pricing tier as GLM-5.1 on this provider; exact parameter count not published by the provider.

Capabilities: Thinking, Chat, Function Calling, Reasoning

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Vision & Document Analysis​

Document Analysis - Small​

  • Context Window: 32,768 tokens
  • Streaming: Supported
  • Availability: API only

Optimized for multilingual dialogue use cases

Capabilities: Document Analysis, Chat, Vision, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Document Analysis - Xtra Small​

  • Context Window: 16,384 tokens
  • Streaming: Supported
  • Availability: API only

Optimized for compact and efficient vision-language model

Capabilities: Document Analysis, Chat, Vision

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Llama 4 Maverick multi modal - Small​

  • Context Window: 32,768 tokens
  • Streaming: Supported
  • Availability: API only

Optimized for text and multimodal experiences

Capabilities: Document Analysis, Chat, Vision, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Document Analysis & OCR - Small (DeepSeek OCR)​

  • Context Window: 8,192 tokens
  • Streaming: Supported
  • Availability: API only

DeepSeek OCR. 3B parameter vision-language model specialized for optical character recognition and document understanding. Excels at converting documents to structured text/markdown, table extraction, and mathematical content recognition.

Capabilities: Document Analysis, Vision

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

inference-miner-u25​

  • Context Window: Not specified
  • Streaming: Supported
  • Availability: API only

Vision-language model optimized for document analysis and parsing.

Capabilities: Vision, Document Analysis

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Multilingual​

Chat, Multi-lingual, Coding & function calling - Small​

  • Context Window: 128,000 tokens
  • Streaming: Supported
  • Availability: Chat UI & API

Mistral

Capabilities: Chat, Multi-lingual, Coding, Function Calling

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Chat & General Purpose​

Search, Chat & Analysis - Small​

  • Context Window: Not specified
  • Streaming: Supported
  • Availability: API only

Optimized for web search and chat. Suitable for artists and content creation, including storytelling

Capabilities: Web Search, Chat, Vision

Pricing:

  • Input: ... per million tokens
  • Output: ... per million tokens

Complete Pricing Overview​

Loading prices...

Pricing Notice

Pricing is subject to change at our discretion.

Cost Estimation​

Task TypeTypical Token UsageRecommended Model
Email or message draft500 input + 300 outputApertus Swiss LLM - Large
10-page document summary10K input + 1K outputChat & Document Analysis & Reasoning - Large
Contract analysis (30 pages)30K input + 2K outputApertus Swiss LLM - Large
Complex reasoning task5K input + 3K outputChat & Document Analysis & Reasoning - Large
Multilingual exchange15K input + 10K outputApertus Swiss LLM - Large

Model Availability​

Schatzi AI offers models across two channels:

  • Chat UI (via chat.schatziai.ch): models available in the Chat UI & API
  • API (via REST API with API keys): models available via API

Chat UI Models​

  • Apertus Swiss LLM - Large
  • Chat, Document Analysis & Agent tasks - Xtra Large
  • Chat, Multi-lingual, Coding & function calling - Small
  • Chat, Document Analysis, Coding & Reasoning - Xtra Large
  • Chat, Vision, Document Analysis & Reasoning - Medium
  • Reasoning & Tool Use - Xtra Large (GLM-5.2)

API-Only Models​

These models are only accessible via API key and are not visible in the Chat UI. See Available API Models for the full API reference.

  • Chat & Document Analysis & Reasoning - Large
  • Document Analysis - Small
  • Document Analysis - Xtra Small
  • Fast Reasoning & Instruction Following - Small
  • Reasoning & Problem Solving - Small
  • Llama 4 Maverick multi modal - Small
  • Reasoning & Agent tasks - Large
  • Reasoning & Problem Solving - Medium
  • Reasoning & Problem Solving - Xtra Large
  • Reasoning & Tool Use - Large (GLM-4.5 Air)
  • Search, Chat & Analysis - Small
  • Document Analysis & OCR - Small (DeepSeek OCR)
  • inference-miner-u25
  • Apertus Swiss LLM - Large (v1.5)

Decommissioned Models​

These models are no longer available and are listed for historical reference.

  • Chat & Document Analysis - Medium — decommissioned on 2026-06-01, replaced by Chat, Multi-lingual, Coding & function calling - Small
  • Search, Chat & Analysis - Large — decommissioned on 2026-06-01, replaced by Chat, Document Analysis, Coding & Reasoning - Xtra Large
  • Apertus Swiss LLM - Small — decommissioned on 2026-05-28, replaced by Apertus Swiss LLM - Large
  • Document Analysis - Medium — decommissioned on 2026-05-28, replaced by Chat & Document Analysis & Reasoning - Large
  • Llama 3.3 Multi-lingual - Medium — decommissioned on 2026-05-28, replaced by Llama 4 Maverick multi modal - Small
  • Chat & Vision - Small (Gemma 3n) — decommissioned on 2026-06-01, replaced by Chat, Vision, Document Analysis, Coding & Reasoning - Medium (Gemma 4)
  • Chat & Function Calling - Small (Granite 3.1) — decommissioned on 2026-06-01, replaced by Chat, Multi-lingual, Coding & function calling - Small
  • Reasoning & Agentic - Large (GPT-OSS 120B) — decommissioned on 2026-05-11, replaced by Chat, Document Analysis, Coding & Reasoning - Xtra Large
  • Embedding - Multilingual (BGE-M3) — decommissioned
  • Embedding - Multilingual (Granite 278M) — decommissioned
  • Chat & Document Analysis - Xtra Xtra Large — decommissioned on 2026-05-28
  • inference-bge-reranker — decommissioned

FAQ​

Q: Can I switch models mid-conversation? A: Yes. You can change models at any time; the conversation context carries over, though very long contexts may be truncated on a model with a smaller context window.

Q: Which model is best for Swiss compliance-sensitive work? A: Apertus Swiss LLM - Large runs on Swiss infrastructure and is optimised for data-sovereignty-sensitive use cases.

Q: Do prices update automatically? A: Yes. Prices shown here are fetched live and always reflect current pricing.