Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.
Context / input limit
Not collected
No model-specific public contract established for this field.
Maximum input tokens
Not collected
No model-specific public contract established for this field.
Maximum output
Not collected
No model-specific public contract established for this field.
Reasoning
Not collected
No model-specific public contract established for this field.
Tool calling
Not collected
No model-specific public contract established for this field.
Structured output
Not collected
No model-specific public contract established for this field.
Streaming
Not collected
No model-specific public contract established for this field.
Batch processing
Not collected
No model-specific public contract established for this field.
Fine-tuning
Not collected
No model-specific public contract established for this field.
Knowledge cutoff
Not collected
No model-specific public contract established for this field.
Release date
Not collected
No model-specific public contract established for this field.
Parameters
Not collected
No model-specific public contract established for this field.
Architecture
Multimodal Mochi Transformer; the later audio research identifies shared architecture and training paradigm with this avatar video model.Source for Architecture (opens in a new tab)
Languages
Not collected
No model-specific public contract established for this field.
Context caching
Not collected
No model-specific public contract established for this field.
Endpoint pricing has not been collected. Check the provider for applicable rates.
Data & enterprise terms
Retention, training use, regional processing and enterprise terms have not been summarized for this endpoint.
Caching
Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.
01 / INITIAL INPUTProcess & retain
The system processes eligible input and retains reusable state.
02 / LATER REQUESTCheck available state
Reuse depends on matching input, session state and the cache policy.
03 / CACHE HITReuse & generate
Reusing state can reduce repeated work. Any charges follow the billing rules below.
Mirage research
Not verified
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
How it works
Not collected
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
Minimum reusable input
Not collected
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
Lifetime & refresh
Not collected
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
How to enable it
Not collected
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
Usage fields
Not collected
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
Billing rules
Not collected
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
Limits & data boundaries
Not collected
No inference-cache contract or discount established. A reusable avatar, stored voice, job ID or legal permission to cache content does not establish prompt caching.
No verified cache rate for this endpoint. An empty rate is not a zero charge.
Caching reviewed 15 Oct 2026
What the measurements say
No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.
Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.
The record behind the model
Why it is included
First-party attribution identifies a Mirage-developed model. Closed-weight status is inferred from its hosted distribution and proprietary-model ownership terms; the reviewed sources establish no downloadable checkpoint or open-weight license.
