name = "Fugu Ultra"
description = "Quality-first multi-agent model for hard research, analysis, and competitions"
family = "fugu"
release_date = "2026-06-15"
last_updated = "2026-06-15"
attachment = true
reasoning = true
temperature = false
tool_call = true
structured_output = true
open_weights = false

[limit]
context = 1_000_000

[modalities]
input = ["text", "image"]
output = ["text"]

[[links]]
label = "Official model catalog"
url = "https://raw.githubusercontent.com/SakanaAI/fugu/refs/heads/main/configs/files/fugu.json"
type = "docs"

[[benchmarks]]
name = "SWE Bench Pro"
score = 73.7
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "Terminal Bench 2.1"
score = 82.1
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "LiveCodeBench"
score = 93.2
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "LiveCodeBench Pro"
score = 90.8
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "Humanity’s Last Exam"
score = 50.0
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "CharXiv Reasoning"
score = 86.6
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "GPQA Diamond"
score = 95.5
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "SciCode"
score = 58.7
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "τ3 Banking"
score = 20.6
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "Long Context Reasoning"
score = 73.3
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "MRCRv2"
score = 93.6
source = "https://console.sakana.ai/models"

[[benchmarks]]
name = "CTI-REALM"
score = 69.4
source = "https://console.sakana.ai/models"