Tutorial: Run inference with Qwen3 235B A22B-2507 Instruct using W&B Inference

pip install openai wandb weave

# Set environment variables in your notebook
%env WANDB_API_KEY=your-wandb-api-key-here

export WANDB_API_KEY="your-wandb-api-key-here"

import os
# Set environment variables programmatically
os.environ["WANDB_API_KEY] = "your-wandb-api-key-here"

import os
import openai
import weave

PROJECT = "wandb_inference"
weave.init(PROJECT)

client = openai.OpenAI(
    base_url="https://api.inference.wandb.ai/v1",
    api_key=os.getenv("WANDB_API_KEY"),
    project=PROJECT,
    default_headers={
        "OpenAI-Project": "wandb_fc/quickstart_playground"  # replace with your actual team/project
    }
)

resp = client.chat.completions.create(
    model="Qwen/Qwen3-235B-A22B-Instruct-2507",
    messages=[
        {"role": "system", "content": "You are a clear, friendly science explainer."},
        {"role": "user", "content": "In plain language, outline the core ideas behind quantum computing—qubits, superposition, entanglement, and gates. Give one real-world application and one current limitation."}
    ],
    temperature=0.6,
    max_tokens=900,
)

print(resp.choices[0].message.content)

import os
import openai
import weave

PROJECT = "wandb_inference"
weave.init(PROJECT)

client = openai.OpenAI(
    base_url="https://api.inference.wandb.ai/v1",
    api_key=os.getenv("WANDB_API_KEY"),
    project=PROJECT,
    default_headers={
        "OpenAI-Project": "wandb_fc/quickstart_playground"  # replace with your actual team/project
    }    
)

resp = client.chat.completions.create(
    model="Qwen/Qwen3-235B-A22B-Instruct-2507",
    messages=[
        {"role": "system", "content": "You are a creative writing coach with a focus on sci‑fi tone and pacing."},
        {"role": "user", "content": "Draft a tight 600–800 word time‑travel story with strong characters, a twist, and a reflective note on cause/effect. Keep prose lean and vivid."}
    ],
    temperature=0.75,
    top_p=0.9,
    max_tokens=1600,
)
print(resp.choices[0].message.content)


import os
import sys
import openai
import weave


PROJECT = "wandb_inference"
weave.init(PROJECT)

client = openai.OpenAI(
    base_url="https://api.inference.wandb.ai/v1",
    api_key=os.getenv("WANDB_API_KEY"),
    project=PROJECT,
    default_headers={
        "OpenAI-Project": "wandb_fc/quickstart_playground"  # replace with your actual team/project
    }
)

stream = client.chat.completions.create(
    model="Qwen/Qwen3-235B-A22B-Instruct-2507",
    messages=[
        {"role": "system", "content": "You are a space history narrator with a balanced, insightful tone."},
        {"role": "user", "content": "Tell the story of space exploration—from early dreams to present—touching on key missions, innovators, international cooperation, big challenges, and why it matters."}
    ],
    stream=True,
    temperature=0.6,
)

sys.stdout.write("Response: ")
for chunk in stream:
    delta = chunk.choices[0].delta
    if delta and delta.content:
        sys.stdout.write(delta.content)
        sys.stdout.flush()
print()


import io
import requests
import openai
import weave
from pypdf import PdfReader
import os


PROJECT = "wandb_inference"
weave.init(PROJECT)

PDF_URL = "https://docs.aws.amazon.com/pdfs/bedrock-agentcore/latest/devguide/bedrock-agentcore-dg.pdf"
QUESTION = "Summarize how AgentCore's memory architecture functions and when to use it."

client = openai.OpenAI(
    base_url="https://api.inference.wandb.ai/v1",
    api_key=os.getenv("WANDB_API_KEY"),
    project=PROJECT,
    default_headers={
        "OpenAI-Project": "wandb_fc/quickstart_playground"  # replace with your actual team/project
    }
)

r = requests.get(PDF_URL, timeout=60)
r.raise_for_status()

reader = PdfReader(io.BytesIO(r.content))
pages = reader.pages[:100]
text = "\n\n".join(page.extract_text() or "" for page in pages)

doc_snippet = text

prompt = (
    f"Using the provided AWS Bedrock AgentCore doc, answer: {QUESTION}\n\n"
    f"Documentation:\n{doc_snippet}\n\n"
    "Cite quotes where relevant."
)

resp = client.chat.completions.create(
    model="Qwen/Qwen3-235B-A22B-Instruct-2507",
    messages=[
        {"role": "system", "content": "You analyze the given text only. If info is missing, say so."},
        {"role": "user", "content": prompt}
    ],
    temperature=0.25,
    max_tokens=1400,
)

print(resp.choices[0].message.content)


import os
import openai
import weave

PROJECT = "wandb_inference"
weave.init(PROJECT)

client = openai.OpenAI(
    base_url="https://api.inference.wandb.ai/v1",
    api_key=os.getenv("WANDB_API_KEY"),
    project=PROJECT,
    default_headers={
        "OpenAI-Project": "wandb_fc/quickstart_playground"  # replace with your actual team/project
    }
)

code_snippet = """
// JavaScript function with Chinese identifiers and mixed language comments
function 计算总价(商品列表, 折扣率) {
    let 总价 = 0;  // Total price accumulator
    for (const 商品 of 商品列表) {
        总价 += 商品.价格 * 商品.数量  // price * quantity
    }
    const 折扣金额 = 总价 * 折扣率  // discount amount
    return 总价 - 折扣金额  // final price
}

# Python with Chinese docstring and mixed naming
def validate_用户输入(user_data):
    '''
    验证用户输入数据的完整性和有效性
    Validates user input data for completeness and validity
    '''
    required_fields = ['name', 'email', '年龄']
    for field in required_fields:
        if field not in user_data:
            raise ValueError(f"Missing required field: {field}")
    return True
"""

task = (
    "Analyze this code: 1) Explain in English+Chinese 2) Flag issues 3) Suggest improvements 4) Refactor with consistent naming."
)

resp = client.chat.completions.create(
    model="Qwen/Qwen3-235B-A22B-Instruct-2507",
    messages=[
        {"role": "system", "content": "You are a bilingual software architect (EN/中文)."},
        {"role": "user", "content": f"{task}\n\nCode:\n{code_snippet}"}
    ],
    temperature=0.2,
    max_tokens=1100,
)

print(resp.choices[0].message.content)


import openai
import weave
import os


PROJECT = "wandb_inference"
weave.init(PROJECT)

client = openai.OpenAI(
    base_url="https://api.inference.wandb.ai/v1",
    api_key=os.getenv("WANDB_API_KEY"),
    project=PROJECT,
    default_headers={
        "OpenAI-Project": "wandb_fc/quickstart_playground"  # replace with your actual team/project
    }
)

resp = client.chat.completions.create(
    model="Qwen/Qwen3-235B-A22B-Instruct-2507",
    messages=[
        {"role": "system", "content": "You are a pragmatic SaaS strategist focusing on retention and efficient growth."},
        {"role": "user", "content": "Given churn 15%, CAC $150, CLV $800, ARPU $50, 2k customers, $60k marketing—provide: benchmarks check, top 3 problems, action plan, 90‑day timeline, and success KPIs."}
    ],
    temperature=0.65,
    max_tokens=900,
)

print(resp.choices[0].message.content)

import os
import openai
import weave

PROJECT = "wandb_inference"
weave.init(PROJECT)

client = openai.OpenAI(
    base_url="https://api.inference.wandb.ai/v1",
    api_key=os.getenv("WANDB_API_KEY"),
    project=PROJECT,
    default_headers={
        "OpenAI-Project": "wandb_fc/quickstart_playground"  # replace with your actual team/project
    }
)

@weave.op()
def analyze_customer_feedback(feedback_text, sentiment_threshold=0.5):
    """Return sentiment (‑1..+1), themes, risks, and next actions."""
    resp = client.chat.completions.create(
        model="Qwen/Qwen3-235B-A22B-Instruct-2507",
        messages=[
            {"role": "system", "content": "You assess customer feedback for sentiment and themes; be concise and actionable."},
            {"role": "user", "content": f"Analyze: {feedback_text}\n\nThreshold: {sentiment_threshold}. Return: score, themes, concerns, actions."}
        ],
        temperature=0.1,
        max_tokens=500,
    )
    return resp.choices[0].message.content

if __name__ == "__main__":
    out = analyze_customer_feedback(
        "Core features are solid, but the latest UI update slowed things down and added friction.",
        sentiment_threshold=0.3,
    )
    print(out)