NNexithDocs
StatusDashboard ↗
Overview & quickstartModel catalogGuidesAPI referenceIntegrationsOAuth2 / OIDC

Model catalog

Updated: 2026-11-19 · The live response from GET /v1/models is authoritative

Nexith publicly offers only two model SKUs. Internally there are also several models used for orchestration and pricing; they do not appear in /v1/models and cannot be selected by callers.

Choosing a model#

  • Not sure which to use → nexith-core. Use it for chat, writing, summarization, code, and vision understanding; the gateway internally picks a stronger upstream based on difficulty and capability, and callers never need to know.
  • You need an image → nexith-image (quality tier, $0.10/image).

Public models#

nexith-core#

General-purpose text model: chat, writing, summarization, code, and vision understanding. 1M context, supports tool calling and reasoning.

Field Value
Model ID nexith-core (alias nexith-8b)
Model code NX-C1
Input Text, images
Output Text
Context window 1,048,576 tokens
Max output 131,072 tokens
Input price $8.00 / 1M tokens
Cached input $0.80 / 1M tokens (10% of the input price)
Output price $30.00 / 1M tokens
Reasoning Supported (reasoning_effort)
Tool calling Supported (tools / tool_calls)
Streaming Supported
Knowledge cutoff 2026-02
Endpoints /v1/chat/completions, /v1/responses, /v1/messages, /v1/completions (legacy)

nexith-image#

Text-to-image and image editing. Priced per image, fixed 1024×1024.

Field Value
Model ID nexith-image
Model code NX-I1
Input Text, images (when editing)
Output Images
Size 1024×1024 (the size field is accepted; currently fixed to this tier)
Price $0.10 / image
Streaming Not supported (async jobs)
Endpoints POST /v1/images/generations; editing via POST /v1/images/edits (chat JWT only)

Image jobs are asynchronous: submission returns 202 and a job id, and the artifact URL is retrieved by polling /v1/jobs/:id or subscribing to /v1/jobs/:id/events. See the Image guide for the full flow.


Internal models (not callable)#

Name Purpose
nexith-standby Internal pricing anchor for large-context/escalation requests, status=internal
nexith-router Internal orchestration calls such as L1 classification and identity rewriting
nexith-cache-embed Local embeddings for the semantic cache and memory (768 dimensions)
nexith-max Historical internal SKU, is_active=false

These models do not appear in GET /v1/models; specifying one explicitly returns 404 model_not_found.


Deprecated#

Model Deprecation date Replacement
nexith-image-fast 2026-11-19 nexith-image

See Deprecations for the reason and migration notes.