{"version":1,"type":"story","url":"https://digestai.news/story/liquid-ai-releases-d1-ultra-decision-model-at-0-042-per-million-tokens","json":"https://digestai.news/story/liquid-ai-releases-d1-ultra-decision-model-at-0-042-per-million-tokens.json","markdown":"https://digestai.news/story/liquid-ai-releases-d1-ultra-decision-model-at-0-042-per-million-tokens.md","slug":"liquid-ai-releases-d1-ultra-decision-model-at-0-042-per-million-tokens","headline":"Liquid AI releases d1-ultra decision model at $0.042 per million tokens","summary":"Liquid AI has released d1-ultra, a multimodal decision model designed to answer bounded questions with probabilities rather than generating free text. The model processes text and images in a single forward pass, producing zero output tokens, which reduces costs for tasks like classification, routing, and moderation. It is currently available via the Liquid API at a rate of $0.042 per million input tokens.\n\nAccording to Liquid AI, d1-ultra builds on a larger open-weight backbone than its predecessor, d1, with customizations that improve performance on knowledge tasks. The company reports that the model ranks #1 on the Artificial Analysis Decision Index with a score of 65, outperforming TypeSafe’s Jev 1.14 across all 112 evaluations in their internal suite. However, specific parameter counts and context lengths have not been disclosed, and the model is only accessible through the hosted API, with no downloadable weights or quantized variants announced.\n\nPricing is based solely on input tokens, with images costing 1.5 tokens per 32x32-pixel patch. Developers can test the model in the d1 Playground, and support for Vercel AI Gateway and OpenRouter is expected soon. The release targets use cases where precise probability outputs are needed, such as spotting defects in circuit boards or verifying insurance claims against photos.","keyPoints":["Liquid AI released d1-ultra, a multimodal decision model returning probabilities with zero output tokens.","The model is available on the Liquid API for $0.042 per million input tokens.","Liquid reports d1-ultra ranks #1 on the Artificial Analysis Decision Index with a score of 65."],"whyItMatters":"d1-ultra offers a cost-effective alternative to large language models for classification and routing tasks by eliminating output token costs. Its high accuracy on decision benchmarks makes it suitable for automated moderation and triage systems.","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"entities":{"companies":["Liquid AI","TypeSafe","Vercel","OpenRouter","Microsoft"],"models":["d1-ultra","d1","Jev 1.14","Jev 1.13","GLiDE","Microsoft-Decision-1"],"people":[]},"firstPublishedAt":"2026-10-16T06:26:41Z","updatedAt":"2026-10-16T06:26:41Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"MarkTechPost","title":"Liquid AI Releases d1-ultra: A Multimodal Decision Model That Scores 65 on the AA Decision Index","url":"https://marktechpost.com/2026/10/15/liquid-ai-d1-ultra-decision-model","publishedAt":"2026-10-16T06:26:41Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":{"title":"The AI Decision Models Competition Saga","url":"https://digestai.news/thread/researchers-introduce-context-language-models-that-manage-their-own-context","storyCount":4},"cite":{"text":"Digest AI, \"Liquid AI releases d1-ultra decision model at $0.042 per million tokens\", 16 October 2026, https://digestai.news/story/liquid-ai-releases-d1-ultra-decision-model-at-0-042-per-million-tokens","publisher":"Digest AI","title":"Liquid AI releases d1-ultra decision model at $0.042 per million tokens","datePublished":"2026-10-16T06:26:41Z","url":"https://digestai.news/story/liquid-ai-releases-d1-ultra-decision-model-at-0-042-per-million-tokens"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}