from openai import OpenAIimport osclient = OpenAI( api_key=os.getenv("API_KEY"), base_url="https://api.friendli.ai/serverless/v1",)completion = client.chat.completions.create( model="Qwen/Qwen3-235B-A22B-Instruct-2507", extra_body={}, messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Tell me a funny joke."}, ],)print(completion.choices[0].message.content)
from openai import OpenAIimport osclient = OpenAI( api_key=os.getenv("API_KEY"), base_url="https://api.friendli.ai/serverless/v1",)completion = client.chat.completions.create( model="Qwen/Qwen3-235B-A22B-Instruct-2507", extra_body={}, messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Tell me a funny joke."}, ],)print(completion.choices[0].message.content)
Qwen3-235B-A22B-Instruct-2507 is an FP8 MoE model from Alibaba's Qwen team with 235B parameters (22B active) in non-thinking mode. It improves tool use, coding, instruction following, logical reasoning, and text comprehension. Native context is 262K tokens, extendable to 1M via Dual Chunk Attention. Apache 2.0 licensed. Strong on math and reasoning, but production deployment needs an optimized inference stack to control latency.
