Resolve the default for empty drafts without persisting a session.

GET

/v1/models/default

const url = 'https://app.everruns.com/api/v1/models/default';
const options = {method: 'GET'};

try {
  const response = await fetch(url, options);
  const data = await response.json();
  console.log(data);
} catch (error) {
  console.error(error);
}

curl --request GET \
  --url https://app.everruns.com/api/v1/models/default

Responses

200

Effective organization default, or null when unavailable

application/json

One of:

object
null

LLM Model with provider info

object

capabilities

required

Capability tags supported by this model.

Array<string>

created_at

required

Timestamp when this model was created (RFC 3339).

string format: date-time

display_name

required

Human-readable display name.

string

enabled

required

Whether this model is selectable. Controls UI visibility AND server-side resolution: ProviderResolverService requires enabled = true, and org default-model validation rejects disabled models.

boolean

healthy

required

Derived: model is configured and ready for use. Currently means the joined provider is active and has an API key set; over time this may also incorporate live reachability checks. Not persisted.

boolean

required

Prefixed public identifier. See ID Schema.

string

is_favorite

required

Whether this model is starred in the UI for quick access.

boolean

model_id

required

Provider-side model identifier as sent on the wire (e.g. gpt-5.2).

string

model_vendor

One of:

string
null

Vendor/brand of the model, derived from the model registry. Drives UI branding (icons). None when the model id is not in the registry. Not persisted.

string

Allowed values: openai anthropic google nvidia qwen microsoft meta minimax moonshot xai llmsim

profile

One of:

object
null

Readonly profile with model capabilities (limits, pricing, modalities). Not persisted.

object

attachment

required

Whether the model supports file/image attachments

boolean

cost

One of:

object
null

Cost per million tokens

object

cache_read

Cached read cost per million tokens (USD), if supported

number | null format: double

cache_write

Cache write cost per million tokens (USD); absent falls back to input.

number | null format: double

cost_tiers

Tiered pricing that applies when prompt tokens exceed context thresholds. When present, the highest matching tier replaces the base rates for the whole request.

Array<object>

A pricing tier that activates above a context token threshold. For example, OpenAI charges higher rates for prompts exceeding 200K tokens.

object

above_tokens

required

Context token threshold above which this tier applies

integer format: int32

cache_read

Cached read cost per million tokens (USD) for this tier, if supported

number | null format: double

cache_write

Cache write cost per million tokens (USD); absent falls back to input.

number | null format: double

input

required

Input cost per million tokens (USD) for this tier

number format: double

output

required

Output cost per million tokens (USD) for this tier

number format: double

input

required

Input cost per million tokens (USD)

number format: double

output

required

Output cost per million tokens (USD)

number format: double

description

Short human-readable description of the model’s strengths and intended use

string | null

family

required

Model family (e.g., “gpt-5.6-sol”, “claude-sonnet-5”)

string

knowledge

Knowledge cutoff date (YYYY-MM-DD format)

string | null

last_updated

Last updated date (YYYY-MM-DD format)

string | null

limits

One of:

object
null

Token limits

object

context

required

Maximum context window size in tokens

integer format: int32

input

Maximum input tokens (if different from context - output)

integer | null format: int32

max_media

Maximum images or PDF pages per request

integer | null format: int32

output

required

Maximum output tokens

integer format: int32

modalities

One of:

object
null

Supported modalities

object

input

required

Supported input modalities

Array<string>

Allowed values: text image audio video pdf

output

required

Supported output modalities

Array<string>

Allowed values: text image audio video pdf

name

required

Display name of the model

string

open_weights

required

Whether the model has open weights

boolean

reasoning

required

Whether the model has reasoning/chain-of-thought capabilities

boolean

reasoning_effort

One of:

object
null

Reasoning effort configuration (for reasoning models)

object

default

required

Default reasoning effort for this model

string

Allowed values: none minimal low medium high xhigh max

values

required

Available reasoning effort values for this model

Array<object>

Named reasoning effort value for UI display

object

name

required

Display name (e.g., “Low”, “Medium”)

string

value

required

The API value (e.g., “low”, “medium”)

string

Allowed values: none minimal low medium high xhigh max

release_date

Release date (YYYY-MM-DD format)

string | null

speed

One of:

object
null

Speed (service tier) configuration, for models served with selectable latency/price tiers (OpenAI service_tier).

object

default

required

Default speed for this model

string

Allowed values: flex default priority fast ultrafast

values

required

Available speed values for this model

Array<object>

Named speed value for UI display

object

cost_multiplier

Price of this tier relative to the standard rate (e.g. 2.0 for Fast, 0.5 for Flex), applied to every token bucket. None when the tier’s rate is not recorded; cost estimates then use the standard rate.

number | null format: double

name

required

Display name (e.g., “Flex”, “Fast”)

string

value

required

The API value (e.g., “flex”, “priority”)

string

Allowed values: flex default priority fast ultrafast

structured_output

required

Whether the model supports structured output (JSON mode)

boolean

supported_parameters

Provider-advertised request parameters supported by this model.

Array<string>

supports_phases

Whether the model supports native execution phases (“commentary” / “final_answer”). When true, the driver sends the phase field on assistant messages in the wire format. Currently supported by GPT-5.4 and newer via OpenAI Responses API.

boolean

supports_server_compaction

Whether the direct provider API supports threshold server-side compaction.

This is an explicit rollout control. Unknown models and provider surfaces that do not opt in remain disabled.

boolean

temperature

required

Whether temperature control is supported

boolean

tool_call

required

Whether the model supports tool/function calling

boolean

tool_search

Whether the model supports tool_search (deferred tool loading). When true, the driver can use namespaces and defer_loading to reduce token usage for large tool sets. Currently supported by GPT-5.4 and newer.

boolean

verbosity

One of:

object
null

Verbosity configuration, for models that expose output-length control (OpenAI verbosity).

object

default

required

Default verbosity for this model

string

Allowed values: low medium high

values

required

Available verbosity values for this model

Array<object>

Named verbosity value for UI display

object

name

required

Display name (e.g., “Low”, “High”)

string

value

required

The API value (e.g., “low”, “high”)

string

Allowed values: low medium high

provider_id

required

Owning provider’s prefixed public identifier.

string

provider_name

required

Joined provider display name.

string

provider_type

required

Joined provider implementation type.

string

source

required

How this model entry was added (manually, discovered, or seeded as predefined).

string

Allowed values: manual discovered predefined

updated_at

required

Timestamp when this model was last updated (RFC 3339).

string format: date-time

Example

{
  "capabilities": [
    "text",
    "tools",
    "vision",
    "thinking"
  ],
  "created_at": "2026-01-04T11:23:00Z",
  "display_name": "Claude Sonnet 5.5",
  "enabled": true,
  "healthy": true,
  "id": "model_01933b5a00007000800000000000001",
  "is_favorite": true,
  "model_id": "claude-sonnet-5-5",
  "model_vendor": "openai",
  "profile": {
    "modalities": {
      "input": [
        "text"
      ],
      "output": [
        "text"
      ]
    },
    "reasoning_effort": {
      "default": "none",
      "values": [
        {
          "value": "none"
        }
      ]
    },
    "speed": {
      "default": "flex",
      "values": [
        {
          "value": "flex"
        }
      ]
    },
    "verbosity": {
      "default": "low",
      "values": [
        {
          "value": "low"
        }
      ]
    }
  },
  "provider_id": "provider_01933b5a00007000800000000000001",
  "provider_name": "Anthropic",
  "source": "manual",
  "updated_at": "2026-05-27T15:24:00Z"
}

403

Forbidden

application/json

Standard error response.

Wire shape is RFC 9457 Problem Details: every error response includes title and status, and may include detail, code, allowed_actions, retry_after_seconds, instance, and type. The content type is rewritten to application/problem+json by [problem_json_content_type].

object

allowed_actions

Recovery actions the caller can take next.

Array<object>

Agent-actionable link describing a follow-up the caller can take. Used in two contexts:

Error recovery — ErrorResponse.allowed_actions carries rels like retry, retry-later, unarchive, get-existing so the agent knows the right next call after a 4xx/429.
Entity hypermedia — WithUrls<T>.allowed_actions carries state-aware rels like cancel, events, self, update on the entity itself so the agent can follow links instead of reconstructing routes from prose.

The shape is intentionally identical across both contexts; the closed rel vocabulary documented in knowledge/execution/api-conventions.md distinguishes them.

object

hint

Short, agent-readable hint (e.g. “Shorten ‘name’ to <= 200 chars.”, “Cancel the active turn for this session.”).

string | null

href

Absolute (preferred) or relative URL the caller may invoke directly. Always present on entity hypermedia actions (WithUrls<T>.allowed_actions); optional on error-recovery actions (ErrorResponse.allowed_actions) where the matching operation_id is enough and the URI is implicit from the failed call.

string | null

method

HTTP method to use against href. Required for entity hypermedia actions; usually omitted on error-recovery actions where the same operation is retried with its original method.

string | null

operation_id

OpenAPI operationId the caller should invoke. Lets an MCP client resolve the call without parsing href.

string | null

rel

required

Link relation describing the action. Closed vocabulary documented in knowledge/execution/api-conventions.md — examples: self, cancel, pause, resume, events, retry, retry-later, unarchive, get-existing, delete, update.

string

schema_ref

OpenAPI $ref to the request-body schema, when the action takes one (e.g. #/components/schemas/UpdateSessionRequest). Lets a tool-calling agent fetch the input shape without scanning the whole spec.

string | null

code

Stable, machine-readable error code (snake_case).

string | null

detail

Human-readable explanation specific to this occurrence.

string | null

instance

Request URI for this occurrence.

string | null

retry_after_seconds

Seconds the caller should wait before retrying (429 / transient 503).

integer | null format: int32

status

required

HTTP status code; mirrors the response status line.

integer format: int32

title

required

Short, human-readable summary of the problem (e.g. “Not Found”).

string

type

RFC 9457 problem type URI. Optional; identifies the problem class.

string | null

Example

{
  "allowed_actions": [
    {
      "method": "POST"
    }
  ],
  "code": "session_not_found",
  "detail": "Session session_01933b5a000070008000000000000001 not found in org org_01933b5a000070008000000000000001.",
  "instance": "/v1/sessions/session_01933b5a000070008000000000000001",
  "retry_after_seconds": 30,
  "status": 404,
  "title": "Session not found",
  "type": "https://docs.everruns.com/errors/session_not_found"
}