Several claims remain unverified. A single news roundup reports that Nvidia released Nemotron 3.5 Super VL and smaller edge variants on October 15, but no other source confirms that release, and no reporting supports the "data center soft launch" described in the same summary. All performance figures come from Meta's own evaluations, and no independent benchmark has been published. Pricing for the required Mac hardware has not been confirmed. Apple is expected to release a 512 GB memory product in late October, but its price has not been disclosed. Running the model locally still requires significant computing power.
For attorneys, the release matters because it tests whether frontier-scale models can run on desktop hardware rather than cloud clusters, which bears on cost, data privacy and on-device AI deployment. Clients evaluating open-weight models should check the Muse Model License terms, particularly for commercial use, before deployment. Watch for independent benchmarks, Apple's late-October pricing, and any confirmation of the Nvidia release, which should not be relied on until a second source reports it.