Fugu Ultra v2
Fugu Ultra v2 is a high-performance model developed by Sakana AI that functions as a learned multi-agent orchestration system. It supports text, image, and file inputs and is designed for complex tasks requiring advanced reasoning and tool-calling capabilities.
Fugu Ultra v2 specs and pricing
- Provider: Sakana
- Released: September 11, 2026
- Context window: 1M tokens
- Max output: 128K tokens
- Input: text, images, files
- Output: text
- Price: $5.00 input / $30 output per 1M tokens
- Reasoning: yes
- Tool calling: yes
What Fugu Ultra v2 is good at
- Analyzing large documents and image-based data
- Complex reasoning tasks requiring multi-agent orchestration
- Automated workflows that involve external tool integration
To-dos to try with Fugu Ultra v2
- Upload a 500-page technical manual and extract all safety protocols into a summary table.
- Analyze a set of architectural diagrams and identify potential structural inconsistencies.
- Use tool-calling to query a database and generate a text report based on the retrieved data.
FAQ
- What is the core architecture of Fugu Ultra v2? It is not a monolithic model, but rather a learned multi-agent orchestration system developed by Sakana AI.
- Can this model process visual information? Yes, Fugu Ultra v2 accepts both image and file inputs in addition to standard text.
- How much does Fugu Ultra v2 cost? Fugu Ultra v2 costs $5.00 per million input tokens and $30 per million output tokens through OpenRouter. In todo.is a typical agent step costs about 18.2 credits ($1 = 100 credits).
- What is the context window of Fugu Ultra v2? Fugu Ultra v2 can read up to 1M tokens at once and write up to 128K tokens in one reply.
- Can I use Fugu Ultra v2 without coding? Yes. In todo.is you pick Fugu Ultra v2 as your agent's model and write to-dos in plain words, on the web, the Android app, WhatsApp or Telegram. No API key needed: credits are included, or you can bring your own key.
JavaScript is required to use the todo.is app.