Open App →
Back to News Feed
Digitimes September 22, 2026 By DIGITIMES neutral

Xiaomi open-sources MiMo-V2.6, publishes the bill for scaling reinforcement learning

AI / LLM
<p class="P1" data-start="2998" data-end="3420">Chinese labs have spent the year arguing that reinforcement learning, not pre-training, holds the next gains. Xiaomi has now put a price on that argument. The company released and open-sourced its MiMo-V2.6 series on September 22, after livestreaming the reinforcement learning post-training run that produced it &mdash; publishing step counts, token consumption, and a running cost meter on a public page as the models trained.
Read original article ↗

Related Articles

Hanwha chair calls for stronger defense AI capabilities

Hanwha Group Chairman Kim Seung-youn has urged Hanwha Systems, the group’s IT solutions arm, to strengthen its defense a

Korea Times · September 22, 2026

Uber Technologies files for FTC review of Delivery Hero stake acquisition

The antitrust watchdog said Tuesday it has received an application from Uber Technologies Inc. regarding its plan to acq

Korea Times · September 22, 2026

EU Kids act: The EU’s plan to make the internet safer for kids

The new EU KIDS ACT bans social media for under-13s but allows "mini accounts" for 13-15-year-olds. No infinite scrollin

Euronews · September 22, 2026

As Russian strikes destroy millions of books, Ukraine's literary future hangs in the balance

More than 13 million books, a third of the country’s annual print run, were destroyed in two months this summer, causing

Euronews · September 22, 2026