Sakana: Fugu Ultra v2

Text input Image input File input Text output
Author's Description

Fugu Ultra v2 is the higher-performance model in Sakana AI's Fugu family. Rather than a single monolithic model, Fugu is a learned multi-agent orchestration system: a language model trained to...

Key Specifications
Cost
+$$$$$
Context
1M
Released
Sep 10, 2026
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Tool Choice Include Reasoning Reasoning Tools Structured Outputs
Features

This model supports the following features:

Reasoning Tools Structured Outputs
Performance Summary

Sakana: Fugu Ultra v2, released on September 10, 2026, is positioned as a high-performance model within Sakana AI's Fugu family, characterized by its multi-agent orchestration system. While demonstrating exceptional reliability with a 100% success rate across all benchmarks, its speed performance tends to be slower, ranking in the 18th percentile. Furthermore, the model is positioned at premium pricing levels, ranking in the 1st percentile for cost. Despite its slower response times and higher cost, Fugu Ultra v2 exhibits remarkable accuracy in several critical areas. It achieved perfect scores in Hallucinations, Reasoning, and Ethics, indicating a strong ability to acknowledge uncertainty, solve complex logical problems, and adhere to ethical principles. Its Coding performance is also very strong at 97.0% accuracy (96th percentile). Instruction Following and Email Classification show solid performance at 83.0% and 99.0% accuracy respectively. General Knowledge and Mathematics, while not top-tier, still demonstrate respectable accuracy at 99.5% and 93.0%. The model's key strengths lie in its high accuracy across various cognitive tasks, particularly in areas requiring precision and ethical understanding. Its primary weaknesses are its longer response times and premium pricing, which may impact its suitability for cost-sensitive or real-time applications.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $5
Completion $30
Input Cache Read $0.5
Web Search $10000

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Sakana AI
Sakana AI | sakana/fugu-ultra-v2-20260911 1M $5 / 1M tokens $30 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by sakana