AINewsnow

Thinking Machines open-sources 594B MoE model trained in 68 hours on Blackwell racks

This story is from 2026-11-14. It is preserved in the archive; the latest stories are on the live feed.

Thinking Machines released Inkling 2 Small, a 594B-parameter Mixture-of-Experts model pretrained end-to-end in 4-bit NVFP4 on Blackwell GB300 racks in 68 hours.

Read the full story at AlphaSignal ↗

Timeline · 1 report

  1. 2026-11-14 06:09 · AlphaSignal
    Thinking Machines' Inkling 2 Small Trains a 594B Model in 68 Hours

More stories

  1. Anthropic IPO analysis, ARC Prize tests Inkling 2 Small, Thinking Machines details guardrail bypass — Latent Space
  2. Inkling 2 Small trained in 68 hours on 12T tokens, claims edge over Muse Spark — Latent Space
  3. Introducing GPT-6.2 Astra and Sol — OpenAI News
  4. Google launches Gemini 4 Neon mid-tier model and opens Genie 4 to Gemini API developers — Google DeepMind YouTube
  5. Anthropic debuts on Nasdaq in record IPO, briefly hitting $2 trillion valuation — Anthropic (X)
  6. Google books Blue Origin's New Glenn for 2029 launch of AI server satellites — Google: The Keyword
  7. AWS expands SageMaker AI with data preparation tools and cross-Region inference routing — AWS Machine Learning Blog
  8. Anthropic's stock debuts strongly, but Wall Street indexes fall — The Globe and Mail

Read the Daily Brief: the day's most important AI stories →