GLM 5.3 Flash with its refusals removed from the weights from $0.15 per 1M tokens. DeepSeek V4.1 Flash uncensored lands soon and runs free for a limited time.
POST api.experientiallabs.ai/v1Both run on our own GPUs with zero data retention.
glm-5.3-flash-abliterated$0.15/$0.50per 1M tokens in and out
GLM-5.3 Flash by Z.ai with its refusal direction removed from the weights. Scores within noise of the base model on MMLU.
Freefor a limited time at launch
DeepSeek V4.1 Flash with its refusals removed from the weights. Coming to the same key soon.
Any OpenAI SDK works. Streams, tool calls and reasoning pass through.
curl https://api.experientiallabs.ai/v1/chat/completions \
-H "Authorization: Bearer $EXPLABS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "glm-5.3-flash-abliterated",
"messages": [{ "role": "user", "content": "Map the endpoints in this capture" }]
}'For authorized testing and research.
Agents that probe the systems your clients signed off on.
Study how attacks and malware work to build the defense.
Map undocumented APIs and keep coding when a frontier model refuses.
The same data controls as every model on the gateway.
Both models run only on routes that keep nothing.
Your requests never train a shared model.
Pro is zero-retention by default for the whole organization.
Billing keeps token counts, cost and timing. Never the prompt text.
One key for these and every other model on the gateway. Read the terms before you start.