Inception Labs' Mercury 2 AI Overtakes Google's DiffusionGemma at Its Own Game
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
★ Tier-1 Source
Inception Labs introduced Mercury 2 on Thursday, calling it the world's fastest reasoning language model.
Key facts
- On AIME 2026—built from real American Invitational Mathematics Examination problems and scored as the percentage solved correctly—Mercury 2 hit 90%
- On GPQA, a PhD-level science benchmark scored the same way, the two models nearly tie: Mercury 2 at 77% against DiffusionGemma's 73.2%
- Inception Labs introduced Mercury 2 on Thursday, calling it the world's fastest reasoning language model
- Mercury 2 isn't open weights, so it's API/cloud for now
Summary
Inception Labs' Mercury 2 generates roughly 1,000 tokens per second and scored 90 on the AIME 2026. DiffusionGemma is free and open-weight on Hugging Face. That puts it in the same speed bracket Google would later claim for DiffusionGemma. The team bet on parallel generation years ago, when it was a contrarian idea.