Skip to content

Models / service contract

BrokenGPT model and service specifications

The stable BrokenGPT service contract: public model alias, private open-source model routing, context window, endpoints, streaming, tools, JSON, limits, and pricing links.

UPDATED 01 Aug 20268 MIN READTECHNICAL GUIDE
01

One stable alias fronts a private model mixture.

The public API model ID is broken-one and the product name is BrokenGPT. Requests are privately routed across a mixture of open-source models. Backend checkpoint and provider identities can change and are deliberately excluded from public responses and metadata.

02

Current service-level specifications.

BrokenGPT public service specification
FieldCurrent public contractNote
Model aliasbroken-oneStable public identifier
Model typePrivately routed mixture of open-source modelsCheckpoint identities are private
Context window1,048,576 tokens by defaultLive deployment configuration is shown on the system page
Maximum outputCaller-selectable within remaining contextNo separate fixed public maximum is advertised
StreamingSupportedSSE on chat completions and the supported Responses subset
JSON outputSupported, non-streamedjson_object and JSON Schema validation
Function toolsSupportedClient executes returned tool calls
Image inputNot advertisedCurrent public message content is text
EndpointsChat completions, Responses, batches, token count, usage, modelsEach has a documented subset
Knowledge limitNot published as one fixed dateRouting can change; verify current facts
Rate limitsPlan and key dependentReturned in headers and account controls
PricingInput and output token basedSee the live pricing page
03

Behavior must be evaluated separately from the schema.

The gateway does not add a keyword censorship filter or platform-authored refusal responder. Every valid prompt reaches configured inference, but the model mixture can still produce refusals, errors, hallucinations, incomplete answers, or insecure code.

  • Use a representative evaluation set for your domain.
  • Measure correctness and refusal as separate axes.
  • Record serving date, prompt, parameters, latency, and cost.
  • Re-run regression prompts after meaningful product changes.
04

The API contract is OpenAI-compatible, not provider-equivalent.

Applications using the supported fields may only need to change the base URL, key, and model alias. Other endpoints, multimodal inputs, hosted tools, and exact output behavior should not be assumed. Consult the compatibility matrix and live documentation before deployment.

STRAIGHT ANSWERS

Frequently asked questions

01Which exact model powers BrokenGPT?

BrokenGPT uses a privately routed mixture of open-source models. Individual backend checkpoint and provider identities are not part of the public contract.

02What is the public model ID?

The stable public API alias is broken-one.

03Does BrokenGPT accept images?

Image input is not advertised on the current public chat-completions contract; message content is text.

BUILD WITH THE LIVE CONTRACT

Test one real request end to end.

Create a scoped key, send a representative prompt, and verify output, usage, limits, and error handling before expanding the integration.

Create an account