Historical two-model system using learned per-clip selection between a far-field Qwen adaptation and a Greek-adapted Canary model. Current October API behavior is recorded separately.
Arbiter calibrated on English development pools. Greek uses the far-field Qwen leg alone with forced-language decoding.
Context caching
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
Access & pricing
Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.
No public immutable checkpoint selector or standalone API contract verified for this dated/internal system.
Access
Historical API-served preview described in the September paper. Current base selector has changed internally; separate access to this revision is unverified.
Regions
Original experimental deployment location not specified
Endpoint pricing has not been collected. Check the provider for applicable rates.
Data & enterprise terms
The paper releases no model weights, data or internal tooling. No standalone processing agreement or data-retention contract is established for this paper-only/datetime-specific record; current preview policies are not backfilled.
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.
KIEFER SA · paper-documented system
Not verified
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
How it works
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
Minimum reusable input
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
Lifetime & refresh
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
How to enable it
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
Usage fields
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
Billing rules
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
Limits & data boundaries
Not collected
No model-specific prompt-cache contract verified. Server-side audio segmentation, speech buffers and reused decoders do not establish discounted token caching.
No verified cache rate for this endpoint. An empty rate is not a zero charge.
What the measurements say
September paper · seven cleaned English datasets
KIEFER SA / Sophea authors · provider-reported
4.35% WER
Table 4 mean under forced-English decoding; AMI close-talk, Earnings22 cleaned AA chunked, Gigaspeech cleaned, LS clean/other, SPGISpeech and VoxPopuli clean-accent. Single-run results with no confidence interval; acknowledged Canary AMI training exposure.
Provider private data, far-field Qwen leg only; forced Greek. Reported 95% CI 24.44–27.19. 24 concurrent clients on one B200 is evaluation setup, not a public account quota.
Provider private data; arbiter routed 52 of 719 clips to Canary. Reported 95% CI 13.01–15.57; separate re-decode scored 14.15. Not current October model performance.
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
The provider-authored production paper identifies internal adapted speech models and combined systems and explicitly releases none of their weights, data or tooling. This record covers the specified proprietary adaptation/system; the original Qwen, Whisper and Canary base weights and stock-weight Turbo are not reclassified as closed.
