Liquid AI's d1-ultra tops Artificial Analysis Decision Index with 85% non-hallucination rate
ArtificialAnlys · x · 2026-10-16
Liquid AI released d1-ultra, scoring 65 to take #1 on the Artificial Analysis Decision Index v1.0, ahead of Jev 1.14 (57), GLiDE and Microsoft-Decision-1 (both 54).
- Hallucination drop: 85% non-hallucination rate on AA-Omniscience-Decision vs d1's 2% — more than double the second-best (Microsoft-Decision-1 at 40%). The tradeoff: accuracy fell from 41% to 32% as the model abstains more often
- Speed & price: 93 ms per task (third fastest behind Celeris-1 Decision at 63 ms and Jev 1.14 at 78 ms); priced at $0.042/1M input tokens, input-only billing, $32 per 1M tasks — in line with Celeris-1 and Microsoft-Decision-1, more than double d1/Mercury Decide ($15), less than half of GPT-6 Luna (Decisions) at $79
- Other results: #1 on When2Call tool calling (93% accuracy), tied #1 with GLiDE on WildGuardTest (avg F1 95%), tied #2 on HLE-MultiChoice (15%)
The Decision Index combines four equally weighted evaluations (knowledge/hallucination, reasoning, tool calling, safety classification) for this emerging class of models that pick from fixed options with probabilities in a single call.
Related event: Liquid AI launches d1-ultra decision model, tops AA Decision Index(16 posts)→
More from Models
- Swapping RLM sub-calls for Jev: half the cost, 6x faster at 1M tokens — lateinteraction · 2026-10-16
- Stanford users say OpenAI cut ChatGPT Edu Pro usage 50% mid-contract — DJiafei · 2026-10-16
- MiniCPM-V 4.5 MoE reportedly hits 13+ tokens/s on iPhone with just 3GB memory — heyshrutimishra · 2026-10-16
- IQuest-Q1 writes a 629-line submarine game from one prompt in 2 minutes — Div_pradeep · 2026-10-16
- Trying Grok bot surfaces an unpaid Cursor balance, blocking account merging — haydendevs · 2026-10-16
- TensorFold 1.0.12 runs 120B vision model on 1-2 DGX Sparks, adds Mac image input — HankYeomans · 2026-10-16