2 models
Ranked by Terminal Bench 2.1
Reasoning Tools JSON 82.1

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

sakana/fugu-ultra 2026-06-15 1M context $5/M input $30/M output
11 providers
Reasoning Tools JSON 80.2

Multi-agent model for routing expert agents across complex analytical tasks

sakana/fugu 2026-06-15 1M context Input not listed Output not listed
1 provider