Skip to content

Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.

fix(llms): map default Claude Sonnet 4.6 to its 1M context window (#7125)

Vidit Ostwal704db1d66fe6
diff --git a/lib/crewai/src/crewai/llms/providers/anthropic/completion.py b/lib/crewai/src/crewai/llms/providers/anthropic/completion.py
index e2afec133..e6bc4255f 100644
--- a/lib/crewai/src/crewai/llms/providers/anthropic/completion.py
+++ b/lib/crewai/src/crewai/llms/providers/anthropic/completion.py
@@ -1953,16 +1953,19 @@ class AnthropicCompletion(BaseLLM):
"""Get the context window size for the model."""
from crewai.llm import CONTEXT_WINDOW_USAGE_RATIO
+ # Current offered models. Unknown and retired IDs fall back to 200k.
context_windows = {
- "claude-3-5-sonnet": 200000,
- "claude-3-5-haiku": 200000,
- "claude-3-opus": 200000,
- "claude-3-sonnet": 200000,
- "claude-3-haiku": 200000,
- "claude-3-7-sonnet": 200000,
- "claude-2.1": 200000,
- "claude-2": 100000,
- "claude-instant": 100000,
+ "claude-fable-5": 1000000,
+ "claude-mythos-5": 1000000,
+ "claude-opus-5": 1000000,
+ "claude-sonnet-5": 1000000,
+ "claude-opus-4-8": 1000000,
+ "claude-opus-4-7": 1000000,
+ "claude-opus-4-6": 1000000,
+ "claude-sonnet-4-6": 1000000,
+ "claude-opus-4-5": 200000,
+ "claude-sonnet-4-5": 200000,
+ "claude-haiku-4-5": 200000,
}
for model_prefix, size in context_windows.items():
diff --git a/lib/crewai/tests/llms/anthropic/test_anthropic.py b/lib/crewai/tests/llms/anthropic/test_anthropic.py
index c26475010..f177a2bc4 100644
--- a/lib/crewai/tests/llms/anthropic/test_anthropic.py
+++ b/lib/crewai/tests/llms/anthropic/test_anthropic.py
@@ -4,7 +4,7 @@ import types
from unittest.mock import AsyncMock, patch, MagicMock
import pytest
-from crewai.llm import LLM
+from crewai.llm import CONTEXT_WINDOW_USAGE_RATIO, LLM
from crewai.crew import Crew
from crewai.agent import Agent
from crewai.task import Task
@@ -116,6 +116,44 @@ def test_anthropic_completion_defaults_to_sonnet_4_6():
assert llm.max_tokens == 128000
+def test_default_anthropic_completion_uses_sonnet_4_6_context_window():
+ from crewai.llms.providers.anthropic.completion import AnthropicCompletion
+
+ llm = AnthropicCompletion()
+ assert llm.get_context_window_size() == int(1_000_000 * CONTEXT_WINDOW_USAGE_RATIO)
+
+
+ "model",
+ [
+ "claude-fable-5",
+ "claude-mythos-5",
+ "claude-opus-5",
+ "claude-sonnet-5",
+ "claude-opus-4-8",
+ "claude-opus-4-7",
+ "claude-opus-4-6",
+ "claude-sonnet-4-6",
+ ],
+)
+def test_current_1m_anthropic_models_use_1m_context_window(model: str) -> None:
+ from crewai.llms.providers.anthropic.completion import AnthropicCompletion
+
+ llm = AnthropicCompletion(model=model)
+ assert llm.get_context_window_size() == int(1_000_000 * CONTEXT_WINDOW_USAGE_RATIO)
+
+
+ "model",
+ ["claude-opus-4-5", "claude-sonnet-4-5", "claude-haiku-4-5"],
+)
+def test_claude_4_5_family_uses_200k_context_window(model: str) -> None:
+ from crewai.llms.providers.anthropic.completion import AnthropicCompletion
+
+ llm = AnthropicCompletion(model=model)
+ assert llm.get_context_window_size() == int(200000 * CONTEXT_WINDOW_USAGE_RATIO)
+
+
def test_anthropic_defaults_max_tokens_to_model_limit():
"""Native Anthropic used to default to 4096, which truncates large tool calls."""
llm = LLM(model="anthropic/claude-unknown-future")
@@ -500,11 +538,10 @@ def test_anthropic_context_window_size():
"""
Test that Anthropic models return correct context window sizes
"""
- llm = LLM(model="anthropic/claude-3-5-sonnet-20241022")
+ llm = LLM(model="anthropic/claude-haiku-4-5")
context_size = llm.get_context_window_size()
- assert context_size > 100000
- assert context_size <= 200000 # But not exceed the actual limit
+ assert context_size == int(200000 * CONTEXT_WINDOW_USAGE_RATIO)
def test_anthropic_message_formatting():