Skip to content

Anthropic

minisweagent.models.anthropic

AnthropicModel 

AnthropicModel(**kwargs)

Bases: LitellmModel

For the use of anthropic models, we need to add explicit cache control marks to the messages or we lose out on the benefits of the cache. Because break points are limited per key, we also need to rotate between different keys if running with multiple agents in parallel threads.

Source code in src/minisweagent/models/litellm_model.py 
26
27
28
29
def __init__(self, **kwargs):
    self.config = LitellmModelConfig(**kwargs)
    self.cost = 0.0
    self.n_calls = 0

query 

query(messages: list[dict], **kwargs) -> dict
Source code in src/minisweagent/models/anthropic.py 
15
16
17
18
19
def query(self, messages: list[dict], **kwargs) -> dict:
    api_key = None
    if rotating_keys := os.getenv("ANTHROPIC_API_KEYS"):
        api_key = get_key_per_thread(rotating_keys.split("::"))
    return super().query(set_cache_control(messages), api_key=api_key, **kwargs)
bug_report Something broken/unclear?

Open an issue on GitHub!
help Open-ended discussions

Join our Slack!
Our projects
SWE-agent SWE-agent SWE-rex SWE-ReX SWE-smith SWE-smith SWE-bench SWE-bench sb-cli sb-cli