NNexithDocs
StatusDashboard ↗
Overview & quickstartModel catalogGuidesAPI referenceIntegrationsOAuth2 / OIDC

Deprecations

Updated: 2026-11-19

This page records models and endpoints that are, or will be, deprecated, along with migration paths. After deprecation, requests return 404 model_not_found (models) or 404 not_found (endpoints).


2026-11-19 · nexith-image-fast deprecation#

Item Details
Status Deprecated (model is_active=false, all routes disabled)
Previous price $0.04 / image
Previous upstream Gemini conversational image generation (via gemini-relay)
Reason The upstream had been unavailable for a long time: 6 job failures over 30 consecutive days (500 do_request_failed), with the last success on 2026-09-04; probing directly with a real key, even plain-text Gemini models returned 500 while /v1/models was fine — a fault in the relay's own upstream channel that cannot be fixed on the platform side
Migration Switch to nexith-image ($0.10/image, quality tier, 40–90s/image)
Frontend "Image Fast" has been removed from the chat site dropdown

Migration example

- {"model": "nexith-image-fast", "prompt": "..."}
+ {"model": "nexith-image",      "prompt": "..."}

Note: nexith-image-fast never delivered a "faster" wall-clock latency (historical measurements were about 3.7 minutes, slower than the quality tier); it was positioned purely as a lower-priced draft tier. If a usable low-cost image upstream becomes available in the future, it will return with a new announcement.


Earlier deprecations#

Object Date Notes
gpu-ornith local GPU route 2026-10-20 Local Ornith inference was disabled in production; nexith-core now runs on third-party upstreams
nexith-max — Internal SKU, is_active=false, never publicly released