Get started
Models
Available models
| Model | ID | Refusals | Vision | Context |
|---|---|---|---|---|
 GLM 5.3 Flash | offmem-glm-5.3-flash-abliterated | Removed | Yes | 1M |
 DeepSeek V4.1 Flash | offmem-deepseek-v4.1-flash-abliterated | Removed | Yes | 1M |
| GLM 5.3 | offmem-glm-5.3-abliterated | Removed | No | 1M |
 Kimi K3.1 Pro | offmem-kimi-k3.1-pro | Original weights | Yes | 1M |
 MiniMax M3.1 Flash | offmem-minimax-m3.1-flash | Original weights | Yes | 1M |
| DeepSeek V4.1 Flash Fast | offmem-deepseek-v4.1-flash-fast | Original weights | Yes | 1M |
- Removed: the refusal behavior was removed from the model (abliterated).
- Original weights: the model as released by its maker; it may refuse some requests.
Prices
In reais (BRL) per 1 million tokens.
| Model | Input | Cached input | Output |
|---|---|---|---|
| GLM 5.3 Flash | R$1.25 | R$0.40 | R$5.15 |
| DeepSeek V4.1 Flash | R$1.45 | R$0.35 | R$5.15 |
| GLM 5.3 | R$10.25 | R$1.55 | R$20.50 |
| Kimi K3.1 Pro | R$2.60 | R$1.55 | R$25.65 |
| MiniMax M3.1 Flash | R$0.55 | R$0.05 | R$2.60 |
| DeepSeek V4.1 Flash Fast | R$0.55 | R$0.05 | R$1.55 |
How the charge is calculated: Billing.
Accepted IDs
Besides the IDs above, the model field accepts:
- The ID without the
offmem-prefix:glm-5.3-flash-abliterated. - The ID with the
offmem/prefix:offmem/offmem-glm-5.3-flash-abliterated. - Names starting with
claude-, for tools made for the Anthropic API: names containinghaikugo tooffmem-glm-5.3-flash-abliterated, all others tooffmem-glm-5.3-abliterated.
An unknown model returns 404 modelo with the list of available IDs. The response always shows the offmem ID in model.
Vision
Models that accept images: offmem-glm-5.3-flash-abliterated, offmem-deepseek-v4.1-flash-abliterated, offmem-kimi-k3.1-pro, offmem-minimax-m3.1-flash, offmem-deepseek-v4.1-flash-fast. Images sent to other models return 400 sem_visao.
cURL
curl https://api.offmem.app/v1/chat/completions \
-H "Authorization: Bearer $OFFMEM_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "offmem-glm-5.3-flash-abliterated",
"messages": [{
"role": "user",
"content": [
{"type": "text", "text": "Describe this image."},
{"type": "image_url", "image_url": {"url": "data:image/png;base64,iVBORw0KG..."}}
]
}]
}'The whole request body, images included, is limited to 4 MB. See Errors and limits.
List models via API
GET /v1/models returns the same IDs in the OpenAI format.