Liquid AI d1.2 and d1.2-ultra: decision models that now hear audio
Liquid AI's decision models add audio input and five new typed tasks - reading values from documents, locating objects in images, estimating counts - with d1.2-ultra topping the Artificial Analysis Decision Index at 75.
Why it matters
Decision models return calibrated probabilities instead of text, with zero output tokens; d1.2-ultra now leads the independent Decision Index, and audio completes the input set alongside text and images.
What you could build with it
A compliance team could pipe thousands of contracts through d1.2's read and extract tasks to pull renewal dates, liability caps and termination clauses as calibrated probabilities, sending only low-confidence documents to a human reviewer.
Does it hold up?
Live on the Liquid API today at $0.02/$0.042 per 1M input tokens with no output-token charge, and d1.2-ultra leads the independent Decision Index - but the five new tasks have no public benchmark yet and weights are not released, so it is hosted-only for now.
Built with Liquid AI d1.2 and d1.2-ultra
- Liquid AI Releases d1.2 and d1.2-ultra: Omnimodal Decision Models That Top the AA Decision Indexmarktechpost · Launch coverage: model IDs, the five new answer types, $0.02/$0.042 per 1M input tokens, and d1.2-ultra's 0.8-point calibration error.
- Liquid AI d1.2: Omnimodal Decision Models Top the Benchmarkbyteiota · Why the timing matters: Google, Microsoft and Cloudflare all shipped competing decision models in 2026, and Liquid just outscored them on the index.
- Liquid AI releases d1.2 decision models with audio inputoossa · Plain-language rundown of the five new tasks and the caveat that the new-task accuracy figures come from Liquid's own tests, not a public benchmark.
Learn more
First spotted on article: source.
More AI models
VoiceStudio: open-source, fully-local ElevenLabs alternative for voice cloning, dubbing, and audiobooksdebpalash/VoiceStudio is a desktop app that runs voice cloning, voice design, video dubbing, dictation, transcription…models · JEV 0.75Edge0-145B: open 145B MoE that runs on a single consumer GPUEdge0 open-sourced 145B and 120B latent-MoE models derived from Qwen3.8-Flash-Next, plus an inference framework that…models · JEV 0.74Strata: run a 125B model on a gaming PCFree, MIT-licensed local inference engine that runs Qwen3.8-Flash-Next (125B MoE, about 6B active per token) on…models · JEV 0.73Google AI Edge Gallery: offline speech recognition and synthesis on-deviceGoogle added fully offline speech recognition and text-to-speech to AI Edge Gallery, with ML Kit APIs on Android and…models · JEV 0.73Soup: fine-tune an 8B LLM on a 4 GB laptop GPU from one YAML fileAn open-source CLI that fine-tunes LLMs from a single YAML file, with layer streaming that trains an 8B model on a 4…models · JEV 0.73Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.