CEO Ramin Hasani's MIT-born startup says the model leads its 112-test suite; it returns probabilities for text and image decisions without generating output tokens.
By [Ryan Merket](https://runtimewire.com/author/ryan-merket)
· Published
Primary source: [X](https://x.com/liquidai/status/2110777964489285875)
Why it matters #
Liquid AI is selling a model that outputs a decision and its probability instead of generated prose. The value depends on whether that confidence stays calibrated on real customer workflows well enough to reduce human review.
Liquid AI, led by co-founder and CEO Ramin Hasani (@ramin_m_h), released d1-ultra on October 15th, a multimodal model that takes text, images or both and returns probabilities for defined decisions in one pass. The company is pitching it for tasks such as identifying defects against a product specification, reviewing damage claims from customer photos and checking whether product images match their listings. It charges $0.042 per million input tokens and does not bill for output tokens.…