Multi-model orchestration system targeting complex agentic work. Its release pool excludes Fable 5, Fable 5.1, and GPT-6 Astra. Independent benchmark coverage is pending.
| Benchmark / source | Score |
|---|---|
HLE (text) Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 56.3% |
GDP.pdf Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 34.3% |
Chartography Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 48.3% |
SWEFish Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 71.8% |
DeepSWE (version unspecified) Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 74.3% |
ProgramBench Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 81% |
GPQA Diamond Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 95.5% |
Toolathon Provider reportSakana launch evaluation; Fable 5/5.1 and GPT-6 Astra excluded from model pool | 80.6% |