OutYet.ai

The daily briefing

Mistral's open-weight frontier model enters July early access

Tuesday, July 7, 2026 · what moved in the AI release cycle in those 24 hours.

A rumor-heavy day in the release cycle, with Mistral AI's open-weight model reaching July early access as the clearest confirmed development. Elsewhere, chatter points to Tencent's Hy3 open MoE model reportedly beating GLM-5.1 with fewer active parameters, and to OpenAI serving GPT-5.6 Sol on Cerebras hardware at very high token speeds. Benchmark leaks around Gemini 3.5 Pro and traces of Grok 4.5 round out an unconfirmed but busy rumor mill.

An AI summary of the day’s tracked items. It may contain errors; every item below links to its source.

Top stories

  1. 1

    Mistral AI Targets Frontier Gap With Open-Weight Model Entering July Early Access - Tech Times

    Mistral's open-weight model entering early access is the day's most concrete release news and signals a bid to close the frontier gap with open weights.

    Mistral next-gen (open-weight MoE) · Mistral AI

  2. 2

    OpenAI is launching GPT-5.6 Sol on Cerebras at up to 750 tokens per second in July. Bleys Goodson estimates Sol may be served across 70-100 Cerebras wafers, with roughly one model layer per wafer: around 3T total parameters, 150B active, 70 layers. It suggests OpenAI did not https://t.co/NJ2HdrUB6CRumor

    Reports of GPT-5.6 Sol launching on Cerebras at up to 750 tokens per second would, if accurate, mark a major shift in inference speed and serving infrastructure.

    GPT-5.6 Sol · @kimmonismus

  3. 3

    Tencent released Hy3 today. Beating GLM-5.1 in blind tests while running fewer active params than the models it's competing with 295B MoE, 21B active, 256K context. And it does this on plain GQA. No sparse attention, no MLA. So the efficiency isn't coming from architectural https://t.co/sVlY4ZETmvRumor

    Tencent's Hy3, a 295B MoE with only 21B active parameters, reportedly beating GLM-5.1 in blind tests suggests notable efficiency gains in open models.

    GLM-5.1 · @kimmonismus

Picked by AI from the day’s tracked items; every link goes to the source.

News & rumors added

Every story we’re tracking →