Liquid AI releases d1.2 omni-modal decision models
Understand this story
Liquid AI has released the decision models d1.2 and d1.2-ultra, adding audio input and five tasks: reading values from text or images, extracting text spans, locating image targets, estimating counts, and multiple choice. Both models are live on the Liquid API: d1.2 costs $0.02 per million input tokens and d1.2-ultra costs $0.042, both billed only on input tokens; images are billed at the text rate at 1.5 tokens per 32×32 pixel block, so a 1024×1024 image counts as 1,536 tokens. According to MarkTechPost, the models currently run only on the hosted Liquid API and the weights have not been released; Liquid AI says they will be opened up "at some point"; new organizations get $2 in free credit on Liquid Console. Liquid AI says both models are coming soon to OpenRouter and Vercel. MarkTechPost reports that d1.2-ultra scores 75 on the Artificial Analysis Decision Index v2.1, the highest among decision models, and d1.2 scores 69.
AI overview based on source reports · Updated 15 hr ago
Coverage timeline
Follow the coverage to explore different perspectives.
- MarkTechPostLiquid AI Releases d1.2 and d1.2-ultra: Omnimodal Decision Models That Top the AA Decision Index
Liquid AI released the d1.2 and d1.2-ultra decision models, which return calibrated probabilities over a fixed set of answers rather than generating text, and now accept text, images and audio plus five new answer types (values, spans, boxes, ranged estimates and multi-label choices). d1.2-ultra scores 75 on the Artificial Analysis Decision Index v2.1, the top score among decision models, while d1.2 scores 69; Liquid AI reports 0.8-point calibration error for d1.2-ultra across 13 benchmarks in its own test. The models are hosted on the Liquid API only at $0.042 and $0.02 per 1M input tokens respectively, with no weights or parameter counts released yet.
- Liquid AI BlogFeaturedNOV 16, 2026Modelsd1.2: State-of-the-Art Omnimodal Decision Models
Liquid AI released d1.2 and d1.2-ultra, decision models that add audio input and five new tasks to the API: reading a value, finding spans, locating objects in images, estimating counts with a range, and multi-select choices. d1.2-ultra scores 75 on the Artificial Analysis Decision Index v2.1, the highest on the index and two points ahead of the previous best, and both models are priced at $0.02 for d1.2 and $0.042 per million input tokens for d1.2-ultra. The models report calibration errors of 0.8 points for d1.2-ultra and 1.2 for d1.2 across 13 benchmarks and 22,505 questions.
Activity for this story
- Current comparable activity
- 6
- Peak comparable activity
- 14Nov 17 18:00
- Change in the past 24 hours
- +31%
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.