abliteration.ai now available on MiniRouter

abliteration.ai is now a direct provider on MiniRouter. Both uncensored fine-tunes support up to 32,768 output tokens, including reasoning.

Two models

• Abliterated Model
  Text and images, 262K context. $1.05 input, $0.105 cached input, $3.15 output per 1M tokens.
  abliteration/abliterated-model
• Abliterated Large v2
  Text only, 1M context. $3.15 input, $0.315 cached input, $5.25 output per 1M tokens.
  abliteration/abliterated-model-large-v2

MiniRouter rates: provider cost plus 5%, billed exactly by usage.

APIs and inputs

Call them on /v1/chat/completions, /v1/responses or /v1/messages, including count_tokens. All three support streaming and reasoning. Chat Completions and Responses support JSON schema outputs and application-run tools. On Chat Completions, Abliterated Model also reads up to 4 base64 images per request.

Use automatic tool choice when streaming. Forced tool choice requires a non-streaming request.

Responses are stateless. The provider reports zero content retention by default; MiniRouter’s privacy policy applies separately.

Try it

Add one field to pin a request to abliteration.ai.

curl https://api.minirouter.sh/v1/chat/completions \
  -H "Authorization: Bearer $MINIROUTER_API_KEY" \
  -H 'Content-Type: application/json' \
  --data '{
    "model": "abliteration/abliterated-model",
    "max_tokens": 1024,
    "messages": [{ "role": "user", "content": "Fix this failing test." }],
    "providerOptions": { "gateway": { "only": ["abliteration"] } }
  }'