Sakana Namazu
Neural Network
Sakana Namazu is a reasoning model by Sakana AI, fine-tuned for Japanese language and business scenarios.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 262 144 tokens
- Max answer length
- 65 536 tokens
- Reasoning effort
- none, high
- Capabilities
- reasoning, function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Sakana Namazu?
Namazu is a Sakana AI reasoning model fine-tuned by the Japanese team for the Japanese language and business context. It focuses on precise instruction following and step-by-step task analysis, making it suitable where logic and accuracy are critical. In BotHub, Sakana Namazu is accessible from Russia without a VPN or foreign card, with pay-as-you-go billing for non-expiring tokens. You can access over 250 models in the same window to compare responses, and the unified OpenAI-compatible API allows you to integrate Namazu and switch models later without rewriting code. Chats are encrypted, and history is not saved. Companies get access to contracts, invoices, EDI, and an admin panel with employee limits. It is useful if you work with Japanese partners and documents, prepare business correspondence for the Japanese market, analyze long regulations, build agents requiring a reasoning module, or compare its responses with other models.Frequently asked questions about Sakana Namazu
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
sakana-namazu is a text model from Sakana with reasoning before answering. It is suitable for analyzing long documents, coding, image analysis, and building agents: it accepts text, images, and documents, can call functions, and returns structured responses.
The model generates up to 65,536 tokens per response — enough for a voluminous document, a large code module, or a detailed analysis. The total context window is 262,144 tokens, so the request and response fit along with long source files.
Yes, reasoning before answering is confirmed by our data. The model breaks down the task step-by-step before formulating the result — this helps with math, logic, code debugging, and analyzing long documents where a chain of reasoning is more important than a quick response.
Yes, the model supports function calling and structured JSON output. You describe the available tools, and it selects the appropriate one and returns arguments in the specified schema. In BotHub, this works via a unified OpenAI-compatible API, so your agent or backend connects in the standard way.
Images and documents — yes, the model accepts them as input and responds with text: upload a screenshot, diagram, table, or contract and ask it to analyze the content. We have no data on audio or video input; please check the BotHub interface for details.
There are no specific notes about languages in our data, so we cannot promise anything specific — it is easier to test it with your own task. In BotHub, you can run a couple of typical queries and, if the result is unsatisfactory, switch to another model without rewriting the integration.