Liquid AI Says d1.2-ultra Leads Decision Index v2.1 at 75
Summary
Liquid AI releases d1.2 and d1.2-ultra, decision models that accept audio input and add five tasks, including locating objects. The company says d1.2-ultra scores 75 on the Artificial Analysis Decision Index v2.1, two points ahead of the previous best. Both run on the Liquid API, at $0.02 for d1.2 and $0.042 per million input tokens for d1.2-ultra.
Key Points
- Liquid AI puts d1.2-ultra's calibration error at 0.8 points over 22,505 questions, versus 7.2 for GPT-6 Luna.
- Liquid AI says both models answer 95% to 97% of held-out questions naming a sound or a spoken language.
- Images bill at 1.5 tokens per 32x32-pixel patch, so a 1024x1024 image costs 1,536 tokens.