Pricing

Fugu

Pay as you go

Pay-as-you-go is designed for heavy production workloads that need maximum reliability. Consumption-based tokens are served at higher priority than monthly plan tokens.

Fugu

Active agents	Billing
1 agent	You pay only the standard rate for the specific underlying model.
Multiple agents	We never stack model fees. You are charged a single rate based on the top tier model involved.

Fugu Ultra

Fixed pricing for fugu-ultra-v1.1 and fugu-ultra-v1.0 (previously fugu-ultra-20260615) per 1M tokens.

Token type	Standard price	Context > 272K
Input	$5	$10
Output	$30	$45
Cached input	$0.50	$1.00

Fugu Cyber

Fixed pricing for fugu-cyber-v1.0per 1M tokens. Fugu Cyber is available only through pay-as-you-go billing. Make sure your API key's "Billing mode" is set to "Pay as you go".

Token type	Standard price	Context > 272K
Input	$6	$12
Output	$36	$54
Cached input	$0.60	$1.20

Usage field details

Outside of standard Responses API output, Fugu Ultra returns usage fields that separate user-visible model work from orchestration work. Note that unlike OpenAI, even though the orchestration tokens are stored in token_details fields, they represent real token usage outside of the input and output tokens and will be counted in the final price of the request. The price will be the same as standard input and output tokens.

{
  "usage": {
    "input_tokens": 120,
    "output_tokens": 80,
    "total_tokens": 200,
    "input_tokens_details": {
      "cached_tokens": 0,
      "orchestration_input_tokens": 0,
      "orchestration_input_cached_tokens": 0
    },
    "output_tokens_details": {
      "orchestration_output_tokens": 0
    }
  }
}

Field	Description
`input_tokens`	Tokens from the user input sent to the first model.
`input_tokens_details.cached_tokens`	Cached input tokens for the user input.
`input_tokens_details.orchestration_input_tokens`	Sum of all input tokens used for orchestration.
`input_tokens_details.orchestration_input_cached_tokens`	Cached input tokens from orchestration.
`output_tokens`	Tokens in the final output.
`output_tokens_details.orchestration_output_tokens`	Output tokens from the orchestration.
`total_tokens`	Total token count returned for the request, including orchestration.

Subscription Plan

Subscription plans are best for individuals and everyday hands-on use. Every tier includes both Fugu and Fugu Ultra. Upgrade when you need longer, heavier, or more frequent sessions.

Plan	Price	Best for	Usage
Standard	$20/month	Lightweight daily usage, occasional API calls, small experiments, and trying Fugu in personal workflows.	Baseline allowance
Pro	$100/month	Focused working sessions for regular coding, review, research, and analysis throughout the week.	10x Standard usage
Max	$200/month	Heavy long-running workloads for power users who keep Fugu active across deeper, longer-running tasks.	20x Standard usage

Sakana Namazu

Sakana Namazu is a Japanese-specialized LLM with built-in tools such as web search and code execution. There are no monthly fees or subscription plans — Sakana Namazu is billed purely on a pay-as-you-go (per-token) basis. Built-in tools (web search and code execution) are billed separately.

sakana-namazu-v1.0

Token pricing for sakana-namazu-v1.0 per 1M tokens.

Token type	Price
Input	$0.95
Output	$4.00
Cached input	$0.15

Thinking tokens are billed at the same rate as output tokens.

Tool usage

Tool	Price	Unit
Web search	$7.00	per 1,000 calls (search queries and page body fetches each count as one call)
Code execution	$0.12	per hour (session retained)

Token accounting with built-in tools

When built-in tools such as web search or code execution are enabled, a single API request runs multiple model calls internally — the model calls a tool, receives the result, and reasons again (the "agent loop"). Charges are calculated as follows:

Input tokens — Each internal call re-sends the full prompt, including the conversation history and all tool results so far. Search results, fetched web page content, and code execution output are therefore billed as input tokens on subsequent internal calls. The input token count in the response's usage object is the total across all internal calls.
Cached input tokens — The prompt prefix shared between internal calls is typically served from the prompt cache and reported as cached_tokens (a subset of the input token count). Cached tokens are billed at the discounted cached-input rate, so input cost grows slowly even in long loops.
Output tokens — The total number of tokens generated across all internal calls.
Tool fees — In addition to token charges, web search is billed per call and code execution by sandbox session time (from creation until the response completes), per the table above. Tool invocations appear as items in the response body (such as web_search_call and code_interpreter_call).

cached_tokens is returned as usage.prompt_tokens_details.cached_tokens on the Chat Completions API and usage.input_tokens_details.cached_tokens on the Responses API.