XNews

Google Releases Gemini Decision 1, Scoring 66 on Artificial Analysis Decision Index

Google Releases Gemini Decision 1, Scoring 66 on Artificial Analysis Decision Index

Google released Gemini Decision 1, its first decision model, which scores 66 on the Artificial Analysis Decision Index v2.1, seven points behind leaders Drex Max v1.1 and GLiDE v1.2 at 73. The model leads knowledge and reasoning benchmarks with 66 on AA-Omniscience-Decision and 49 on HLE-MultiChoice, but trails on the AA-Paperclip classification and extraction benchmark. Priced at $0.10 per 1M input tokens, it costs $345 per 1M Tasks, more than three times the $103 cost of Drex Max v1.1.

Event version 1.

Key insights

Mid-pack on Decision Index

Gemini Decision 1 scores 66 on the Artificial Analysis Decision Index v2.1, in line with Elephant 9B and Jev 1.14, and seven points behind leaders Drex Max v1.1 and GLiDE v1.2, which both score 73.

Leads knowledge and reasoning benchmarks

Gemini Decision 1 scores 66 on AA-Omniscience-Decision and 49 on HLE-MultiChoice, leading those benchmarks, but trails on the AA-Paperclip classification and extraction benchmark, answering all 15 questions correctly on 33% of documents versus 63% for Drex Max v1.1.

Cost per task exceeds leading models

Gemini Decision 1 costs $345 per 1M Tasks, compared with $103 for Drex Max v1.1 and $156 for Jev 1.14, driven by its $0.10 per 1M input token price versus $0.03 for Drex Max v1.1; it is similar to GPT-6 Luna at $339.

Strong calibration on reasoning tasks

Gemini Decision 1 has a Brier score of 0.18 across the Index, between Drex Max v1.1 at 0.16 and Jev 1.14 at 0.19, with a 0.20 score on HLE-MultiChoice among the lowest of all models and 0.14 on When2Call versus 0.04 for Drex Max v1.1.

Source posts