redshift c1c6c81108 feat: expose per-model reasoning effort on /v1/models ...

Keep upstream reasoning metadata (supported_efforts, default_effort,
mandatory) on the catalog instead of dropping it on ingest, and map
client reasoning_effort / reasoning.effort / thinking onto each model's
allowlist before forwarding so unsupported levels are not sent upstream.

2026-09-09 13:31:19 +02:00

..

__init__.py

small fixes

2025-11-11 15:42:26 +01:00

cost_calculation.py

feat(payment): keep the upstream reported USD cost beside the billed one

2026-09-02 07:43:24 +02:00

helpers.py

Validate destinations of pre-auth image URL fetches

2026-09-04 02:08:30 +02:00

lnurl.py

fix: address review findings on cooldowns, melt sizing, and stream finalization

2026-08-26 23:15:40 +02:00

models.py

feat: expose per-model reasoning effort on /v1/models

2026-09-09 13:31:19 +02:00

price.py

fix(pricing): reject an exchange quote that underflows the sats price

2026-08-26 15:50:26 +02:00

rates.py

fix(pricing): a rate spelled as a boolean is not a rate

2026-08-26 15:44:29 +02:00

usage.py

revert: remove unrelated repository formatting

2026-07-26 20:12:12 +02:00