Fugu Ultra

sakana/fugu-ultra

multi-agent orchestration across frontier models.

cmd --model sakana/fugu-ultra
Intelligence index
not yet scored
Output speed
not yet scored
Input
$5 /M
Output
$30 /M
Cache read
$0.50 /M
Agent-loop cost
$1.85 /M in
Context window
1M tokens
Released
Modalities

vs. the lineup

Fugu Ultra beside its stablemates and nearest rivals. The ◆ marks the best value in each column across every row shown.

pin a rival:
ModelIntelligenceCodingSpeedInput $/MOutput $/MBlended $/MContext
Claude Opus 563.17852.6$5$25$101M
Step 3.7 Flash30.939.6391.3$0.20$1.15$0.44256K
Claude Haiku 4.524.196.3$1$5$2200K
Fugu Ultra$5$30$11.251M
DeepSeek V4 Pro (latest)pinned45.359.463.3$0.66$1.98$0.991M
DeepSeek V4 Flash (latest)pinned5269.1115.9$0.22$0.66$0.331M

usage calculator

How far a month of credits goes on Fugu Ultra.

Fugu Ultra runs on Max and up — from $100/mo.
Input tokensfresh prompt
800
Output tokensmodel reply
180
Cache read tokensre-read context
50K
cost / request $0.034 · in $5 · out $30 · cache $0.50 per M
fresh input 12%output 16%cache reads 73%
Requests / 30 days
2.9K
$100 credits ÷ $0.034 per request
~581 quick fixes~116 bug fixes~19 feature PRs

what real work costs

Real coding tasks priced end to end on Fugu Ultra, from a quick lookup to a full-repo agent run.

One agent task
$0.65
180K in at 75% cache hit, 12K out
What you pay
$3.40 /M
all-in across every token that task touched
Sticker input
$5 /M
cache reads bill at $0.50 /M instead
TaskTokens in · outFugu UltraClaude Opus 5Claude Haiku 4.5
Quick lookup / one-liner8K · 1K$0.04$0.03$0.0065
Review a 500-line PR60K · 4K$0.23$0.21$0.04
Fix a bug (agent loop)180K · 12K$0.65$0.59$0.12
Refactor a module320K · 20K$1.16$1.06$0.21
Full-repo agent run900K · 45K$2.69$2.47$0.49

frequently asked

What is Fugu Ultra best for?
Fugu Ultra 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 route tasks across a swappable pool of underlying models and to recursively call instances of itself. Fugu Ultra prioritizes answer quality on complex, multi-step reasoning, coding, and agentic workflows. It supports configurable reasoning effort, tool calling, and built-in web search. Orchestration tokens consumed by the system are billed as standard input/output tokens.
How much does Fugu Ultra cost?
$5/M input and $30/M output, cache reads $0.50/M. In an agent loop most input is cache-read, so the effective input rate is about $1.85/M.
Which plan do I need?
Available on Max and above.
How do I switch to it?
Run cmd --model sakana/fugu-ultra, or type /model in a session and pick it. You can switch mid-session without losing context.

Ship code that matches your taste

Command Code is the AI coding agent that continuously learns your taste. Start for $1.

Benchmarks from Artificial Analysis (v4.1)commandcode.ai/models/fugu-ultra