Fimeg/letta-server
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

fix: max output tokens for gemini 3 models (#7322)

Browse Source
This commit is contained in:
Ari Webb
2025-12-17 13:51:58 -08:00
committed by Caren Thomas
parent 25a7eaafe6
commit e47eb23522
4 changed files with 7 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
letta/constants.py
View File

@@ -360,6 +360,9 @@ LLM_MAX_CONTEXT_WINDOW = {
"gemini-2.5-flash-preview-09-2025": 1048576,
"gemini-2.5-flash-lite-preview-09-2025": 1048576,
"gemini-2.5-computer-use-preview-10-2025": 1048576,
# gemini 3
"gemini-3-pro-preview": 1048576,
"gemini-3-flash-preview": 1048576,
# gemini latest aliases
"gemini-flash-latest": 1048576,
"gemini-flash-lite-latest": 1048576,

2
letta/llm_api/google_constants.py
View File

@@ -1,4 +1,6 @@
GOOGLE_MODEL_TO_CONTEXT_LENGTH = {
"gemini-3-pro-preview": 1048576,
"gemini-3-flash-preview": 1048576,
"gemini-2.5-pro": 1048576,
"gemini-2.5-flash": 1048576,
"gemini-live-2.5-flash": 1048576,

2
letta/schemas/providers/google_gemini.py
View File

@@ -28,7 +28,7 @@ class GoogleAIProvider(Provider):
def get_default_max_output_tokens(self, model_name: str) -> int:
"""Get the default max output tokens for Google Gemini models."""
if "2.5" in model_name or "2-5" in model_name: # gemini-2.5-* or gemini-2-5-*
if "2.5" in model_name or "2-5" in model_name or model_name.startswith("gemini-3"):
return 65536
return 8192 # default for google gemini

2
letta/schemas/providers/google_vertex.py
View File

@@ -18,7 +18,7 @@ class GoogleVertexProvider(Provider):
def get_default_max_output_tokens(self, model_name: str) -> int:
"""Get the default max output tokens for Google Vertex models."""
if "2.5" in model_name or "2-5" in model_name: # gemini-2.5-* or gemini-2-5-*
if "2.5" in model_name or "2-5" in model_name or model_name.startswith("gemini-3"):
return 65536
return 8192 # default for google vertex