DeepSeek V4 Flash

DeepSeek V4 Flash is the lightweight, low-latency sibling of V4 Pro. Same 1M-token context window, same reasoning ('thinking') mode, with fewer activated parameters and lower per-token cost. Designed for high-concurrency production traffic, agent loops, and backend chat where the full reasoning depth of Pro isn't required. Function calling and structured output are supported; the model ships FP8-quantised for efficient serving.

Quick start

Use any OpenAI-compatible client. Just change base_url and your key.

from openai import OpenAI

client = OpenAI(
    api_key="tl-xxxxxxxxxxxxxxxxxxxxxxxx",
    base_url="https://api.thalam.ai/v1",
)

response = client.chat.completions.create(
    model="deepseek/deepseek-v4-flash",
    messages=[
        {"role": "user", "content": "Hello from Thalam!"}
    ],
)

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