Try it
Runs in your browser against the live API. Nothing to install.
Reasoning models spend tokens thinking before they answer
A reasoning model uses part of your max_tokens budget on internal reasoning. If the budget is small it can spend ALL of it thinking and return no answer text (the content field comes back empty or null) -- which looks like a broken service but is not. How much each model needs differs a lot; each model's card states the budget measured for it. When in doubt, leave max_tokens unset or switch thinking off where the card says how. The reasoning itself arrives in a separate reasoning_content field; a reply may begin with blank lines, and content can be null when the reply carries a tool call or is cut off at max_tokens, so client code should handle null.
Run the same thing from a terminal
These update as you change the test, key, model and prompt above.curl https://llm.ccs.uky.edu/v1/chat/completions \
-H "Authorization: Bearer YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "ccs/Qwen/Qwen3.6-35B-A3B-FP8-nothink",
"messages": [
{
"role": "user",
"content": "What is 2+2? Answer in one short sentence."
}
],
"max_tokens": 512
}'
pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://llm.ccs.uky.edu/v1",
api_key="YOUR_KEY",
)
resp = client.chat.completions.create(
model="ccs/Qwen/Qwen3.6-35B-A3B-FP8-nothink",
messages=[{"role": "user", "content": "What is 2+2? Answer in one short sentence."}],
max_tokens=512,
)
print(resp.choices[0].message.content)
# Two variables and every OpenAI-compatible tool works: the openai
# CLI, aider, llm, LangChain, LlamaIndex, Continue, Open WebUI.
# Put them in a job script, not in a file you commit.
export OPENAI_BASE_URL="https://llm.ccs.uky.edu/v1"
export OPENAI_API_KEY="YOUR_KEY"
# Check the key works and see what is running right now:
curl -s "$OPENAI_BASE_URL/models" \
-H "Authorization: Bearer $OPENAI_API_KEY"
A copied line goes into your shell history. Prefix it with a space,
or use the Environment tab, if that matters to you. Full base URL:
https://llm.ccs.uky.edu/v1
· no key yet? Get one from
Locksmith
· more clients in the documentation
· what each model is for, on the
status page.