← Model list

Fugu Ultra

sakana·sakana/fugu-ultra·GA·Closed·fugu series·NEW

Quality-first multi-agent model for hard research, analysis, and competitions

At a glance
Official price$5.00 / $30.00 per 1M
Cache read $0.50 · Sakana AI
Lowest paid$4.50 / $27.00
Requesty Gateway · 1.2× spread
Context1,000,000
⚠ Providers report 1,000,000–1,048,576; the table below is authoritative
Output limit1,000,000
Capabilities
ReasoningTool useStructured outputTemperatureAttachments
⚠ Providers report capability flags inconsistently
Modalities
TextImage
Knowledge cutoff
Released / updated2026-06-15 / 2026-06-15

Quality & performance

Artificial Analysis doesn't cover this model (267 of 2059 have data). Quality data comes from independent evals covering widely used models.

Available at 11 providers11 with public prices

ProviderTierInputOutputCache readCache writeContextOutput limitStatus
RequestyGateway$4.50
tiered >272K: $9.00
$27.00$0.451,048,576131,072
OpenRouterGateway$5.00
tiered >272K: $10.00
$30.00$0.501,000,000128,000
LLMTRGateway$5.00$30.001,000,0001,000,000
Merge GatewayGateway$5.00$30.00$0.501,000,000250,000
Sakana AIOfficial
fugu-ultra-20260615
First-party$5.00
tiered >272K: $10.00
$30.00$0.501,000,0001,000,000
Sakana AIOfficialFirst-party$5.00
tiered >272K: $10.00
$30.00$0.501,000,0001,000,000
Vercel AI GatewayCloud$5.00$30.00$0.501,000,0001,000,000
DevPass (LLM Gateway)Gateway$5.00$30.00$0.501,000,0001,000,000
Kilo GatewayGateway$5.00$30.00$0.501,000,000128,000
PioneerGateway$5.00$30.00$0.50$5.001,000,000131,072
NanoGPTGateway$5.25$31.50$0.5251,000,00016,384

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Your usage cost

1Requesty$1,764.00
2OpenRouter$1,960.00
3Merge Gateway$1,960.00
4Sakana AI · Official$1,960.00
5Sakana AI · Official$1,960.00
6Vercel AI Gateway$1,960.00
Switch to Requesty to save $196.00/mo (10%). The gap is small, so staying on the official channel is fine.
Note: this is a gateway; verify availability and rate limits yourself.

Benchmark12 items

NameConditionsScoreMetricSource
SWE Bench Pro73.7Source ↗
Terminal Bench 2.182.1Source ↗
LiveCodeBench93.2Source ↗
LiveCodeBench Pro90.8Source ↗
Humanity’s Last Exam50Source ↗
CharXiv Reasoning86.6Source ↗
GPQA Diamond95.5Source ↗
SciCode58.7Source ↗
τ3 Banking20.6Source ↗
Long Context Reasoning73.3Source ↗
MRCRv293.6Source ↗
CTI-REALM69.4Source ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Reasoning control

effort = higheffort = xhigheffort = maxeffort = loweffort = medium

2 / 11 providers expose no reasoning control (reasoning_options: []).

Related models

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$30.00$02026-08-052026-08-212026-08-05 · Input list $5.002026-08-06 · Input list $5.002026-08-07 · Input list $5.002026-08-08 · Input list $5.002026-08-09 · Input list $5.002026-08-10 · Input list $5.002026-08-11 · Input list $5.002026-08-12 · Input list $5.002026-08-13 · Input list $5.002026-08-21 · Input list $5.002026-08-05 · Output list $30.002026-08-06 · Output list $30.002026-08-07 · Output list $30.002026-08-08 · Output list $30.002026-08-09 · Output list $30.002026-08-10 · Output list $30.002026-08-11 · Output list $30.002026-08-12 · Output list $30.002026-08-13 · Output list $30.002026-08-21 · Output list $30.002026-08-05 · Min blended $11.252026-08-06 · Min blended $11.252026-08-07 · Min blended $11.252026-08-08 · Min blended $11.252026-08-09 · Min blended $11.252026-08-10 · Min blended $11.252026-08-11 · Min blended $11.252026-08-12 · Min blended $11.252026-08-13 · Min blended $11.252026-08-21 · Min blended $10.13