Sakana AI · Fugu
Sakana AI's top hosted multi-model orchestration system (routes and combines frontier models from its pool), aimed at hard coding, research and document tasks via API.
Leader: Claude Opus 5.5 at 61.4%
Leader: GPT-6 Astra at 96.1%
| Result | Score | Reported by | Source |
|---|---|---|---|
| GDP.pdf | 34.3% | vendor | Sakana AI: Fugu Ultra v2 benchmark chart ↗ |
| Chartography | 48.3% | vendor | Sakana AI: Fugu Ultra v2 benchmark chart ↗ |
| SWEFish (Sakana internal) | 71.8% | vendor | Sakana AI: Fugu Ultra v2 benchmark chart ↗ |
| DeepSWE | 74.3% | vendor | Sakana AI: Fugu Ultra v2 benchmark chart ↗ |
| ProgramBench | 81% | vendor | Sakana AI: Fugu Ultra v2 benchmark chart ↗ |
| Toolathlon | 80.6% | vendor | Sakana AI: Fugu Ultra v2 benchmark chart ↗ |