model
Roughly 30B dense multimodal open-weight agent model under Apache 2.0, sized to run on a 24GB consumer GPU.
No evidence-backed relationships are recorded.
Both figures come from the providers' own status pages, and the sharper test is the afternoon of September 3, when OpenAI and Anthropic were degraded inside the same window, which is what a two-subscription fallback has to survive.
Reality
Speaker turns and end-of-speech markers now ride in the same token sequence as the words, which buys a cheaper pipeline and a tighter coupling than a three-vendor speech stack. The accuracy claim behind it is English-only.
Reality
The framework's inputs are token lengths, concurrency, a latency percentile and the share of prompt tokens already sitting in the KV cache, and each of those changes arithmetic that a per-GPU throughput rating leaves untouched.
Reality
The Hatch tier buys capacity, not intelligence, and the alternative it answers runs free on a user's own machine. Whatever Meta finally charges becomes the ceiling everyone else gets priced against.
Reality
The update adds a path selector and a two-tap convolution rather than layers, recovering most of the accuracy that tripling the drafter bought at 15.2% latency, by the vendor's own numbers.
Reality
Muse Glimmer ships as Apache 2.0 weights sized for a 24GB card. Muse Spark 1.2 stays on Muse Code and the Meta Model API. Plan capacity for two tiers, not one.
Perspective Coverage
6 publishersReality