
Meta Releases Muse Glimmer, 30B Parameter Open License Agent Running on 24 GB GPU
Distilled model from Muse Spark arrives under Apache 2.0 on Hugging Face and Ollama. It's Meta's answer to Anthropic and OpenAI's dominance in agentic code.
Strategy
226 analyses

Distilled model from Muse Spark arrives under Apache 2.0 on Hugging Face and Ollama. It's Meta's answer to Anthropic and OpenAI's dominance in agentic code.

Announcement made by Elon Musk during SpaceX's earnings call, detailed on August 6, sidelines AMD, positions Rubin as the brain of the Starmind AI1 satellite, and outlines an off-Earth AI compute system.

Zuckerberg announced on Wednesday the first programming agent from Meta. Terminal-Bench 2.1 positions Muse Spark 1.2 at 82.9%, behind Claude Opus 5 but ahead of GPT-5.6 Terra, priced at $1.25 per million input tokens.

Demis Hassabis leaves the helm of DeepMind to become chairman of the unit and chief scientist of Alphabet. Jeff Dean departs from Google after 27 years to found Discovery Loop, with Alphabet among the investors.

The new multimodal model activates 95 billion parameters per token and processes 1 million context tokens. It leaps straight to the top of Chinese benchmarks, still trailing behind Anthropic's Claude Fable 5.

K&L Gates laid off 10% of its non-legal staff in July, the same week Kirkland accelerated $500 million in proprietary AI. Legal work won't disappear, but the ‘bodega’ billing model for new associates will.

With the European AI Office authorized to fine GPAI since August 2, signing the Code of Practice has become a practical filter. Each instance of non-signature turns into a costly defense case by case.

The three largest banks in continental Europe posted record profits in Q2 2026 as executives attributed measurable efficiency gains to AI for the first time in earnings calls. The cycle of layoffs seen in the U.S. has yet to reach Frankfurt or Zurich.

Launched on July 31, V4-Flash-0731 scores 82.7 on TerminalBench 2.1, compared to 72.1 for V4-Pro-Preview, maintaining the same architecture as the April preview and costing $0.14 per million input tokens.

New model controls legs, torso, arms, and hands with 22 degrees of freedom under a single learned policy. Real deployment begins in the coming months with four partner manufacturers.

AWS revenue reached $42.2 billion at the fastest pace in five years. Andy Jassy raised the 2026 capex to $220 billion and admitted it still won't be enough to meet demand.

Azure's annual revenue surpassed $100 billion for the first time, with a 43% jump in the quarter. Amy Hood guided the market to an estimated capex of approximately $190 billion for the calendar year.