Models

Supported AI models

Onysoft AI Gateway offers models from the world's leading AI providers through a single API. All models can be used in an OpenAI-compatible format.

Available Models

Model	Provider	Context	Input ($/M)
`anthropic/claude-fable-5`	Anthropic	1M	$15.00
`anthropic/claude-sonnet-5`	Anthropic	1M	$3.00
`openai/gpt-5.5`	OpenAI	1M	$7.50
`openai/gpt-5.4-mini`	OpenAI	400K	$1.13
`google/gemini-3.1-pro-preview`	Google	1M	$3.00
`google/gemini-3.5-flash`	Google	1M	$2.25
`meta-llama/llama-4-maverick`	Meta	1M	$0.23
`deepseek/deepseek-v3.2`	DeepSeek	131K	$0.34

View all models and current prices →

Free Türkiye Models (:tr)

Models whose id ends with ":tr" are hosted in Türkiye and are free: no amount is deducted from your balance for these requests. They use the same API key, the same endpoints and the same request body; only the model id differs.

Model	Capability	Endpoint
`z-ai/glm-5.3:tr`	Chat	`/v1/chat/completions`
`qwen/qwen3.8-flash-next:tr`	Chat	`/v1/chat/completions`
`google/gemma-4-31b:tr`	Chat (multimodal)	`/v1/chat/completions`
`qwen/qwen3-vl-30b:tr`	Vision	`/v1/chat/completions`
`qwen/qwen3-embedding-8b:tr`	Embeddings	`/v1/embeddings`
`qwen/qwen3-reranker-8b:tr`	Reranking	`/v1/rerank`

Limits

100 requests per day per API key.
Input size per request is capped; larger requests are rejected with a clear error.
Sending many requests back to back may hit a short burst limit; the system retries once automatically, and if it is still busy, wait a few seconds and try again.
Because they are free they appear in your usage log with an amount of 0.
The models that accept images are google/gemma-4-31b:tr and qwen/qwen3-vl-30b:tr; the others are text only. The vision models can get busy under back-to-back requests, so leaving a few seconds between calls is enough.

curl https://api.onysoft.com/v1/chat/completions \
  -H "Authorization: Bearer sk-ony-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-5.3:tr",
    "messages": [{"role": "user", "content": "Merhaba"}]
  }'

Being free, these models can be busy depending on capacity. For production work that must not be interrupted, prefer the paid models in the catalog.

Typed Decision Endpoint (/v1/decisions)

This endpoint does not produce text. It returns a calibrated probability between 0 and 1 for the binary question you ask — "does this customer want a human?", "is this message reporting a blocking outage?". It is faster and cheaper than forcing a JSON schema on a chat model and parsing the answer, and parse errors do not exist here, because what comes back is already a number.

For detailed usage, question types and error codes, see the Typed Decisions API section.