Anthropic and Responses APIs
Both formats reach the same models, with the same prices and limits as /chat/completions.
What each format supports
| Feature | Chat Completions | Responses | Anthropic Messages |
|---|---|---|---|
| Endpoint | /chat/completions | /responses | /v1/messages |
| Streaming | Yes | Yes | Yes |
| Tool calling | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| Video | Yes | Dropped, still billed | No (400) |
| Reasoning effort | Yes | Yes | Yes |
| Thinking text | Yes | Yes | With thinking on |
| Prompt caching | Yes | Yes | Yes |
| previous_response_id | — | No | — |
| PDFs and token counting | — | — | No |
| Hosted tools such as web search | No | No | No |
Responses stores nothing: send the whole conversation in input with store set to false.
Examples
import os
from anthropic import Anthropic
client = Anthropic(
base_url="https://api.silkcompute.com",
api_key=os.environ["SILK_API_KEY"],
)
message = client.messages.create(
model="deepseek-v4.1-flash-1017-uncensored",
max_tokens=1024,
messages=[{"role": "user", "content": "Explain why the sky is blue."}],
)
print(message.content[0].text)Errors
/v1/messages returns errors in the Anthropic shape, {"type": "error", "error": {"type", "message"}}, with the same status codes as the other endpoints.