Docs/Inference/Models

Models

Two launch models. The table below is read live from the catalog, so prices here are always the prices you are billed.

GET/v1/models
2 models
read live from the catalogbars: output price, relative

Catalog

Model idContextMax output$ in / 1M$ out / 1MCapabilities
venym/qwen-27b
Qwen 27B Unfiltered
131.1K32.8K$0.30$1.20chat / tools / json / reasoning / web
venym/glm-flash
GLM 5.3 Flash Unfiltered
131.1K16.4K$0.10$0.40chat / tools / json / web

Context and max output are in tokens. max_tokensis capped at the model's max output, so asking for more is not an error, you just get the cap.

Capabilities

chatchat completions
All models.
toolsfunction calling
See tools.
jsonJSON output
Pass response_format. See chat completions.
reasoningthinks before answering
The engine may stream reasoning text as part of the response.
webweb grounding
Works with the :online suffix. See web search.

Picking one

  • venym/qwen-27b when quality matters: reasoning, code, long-form writing.
  • venym/glm-flash when speed or volume matters: roleplay, agents, bulk jobs.

You can also pass the short name (qwen-27b) or the display name, case-insensitive. Use the full id in code anyway, it is the one that is returned in responses.

GET /v1/models

Public, no key needed. Returns an OpenAI-style list of the enabled models. pricing is USD per token.

GET /v1/modelscURL
curl https://venym.io/v1/models
payload.jsonJSON
{  "object": "list",  "data": [    {      "id": "venym/qwen-27b",      "object": "model",      "created": 1767225600,      "owned_by": "venym",      "name": "Qwen 27B Unfiltered",      "context_length": 131072,      "max_output_tokens": 32768,      "pricing": { "prompt": 3e-7, "completion": 1.2e-6, "unit": "usd_per_token" },      "capabilities": ["chat", "tools", "json", "reasoning", "web"]    }  ]}