scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-08 22:21:35 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 78

Support no reasoning option for gemini models (#11393)

Browse source 
* support no reasoning for gemini models

* change none to disable

* remove print statements

* update docs
This commit is contained in:
Low Jian Sheng 2025-06-05 15:11:45 +08:00 • committed by GitHub
parent 505d2fe0c7
commit a3e5bc4856
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 15 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
docs/my-website/docs/providers/gemini.md
View file

@ -63,10 +63,13 @@ response = completion(
LiteLLM translates OpenAI's `reasoning_effort` to Gemini's `thinking` parameter. [Code](https://github.com/BerriAI/litellm/blob/620664921902d7a9bfb29897a7b27c1a7ef4ddfb/litellm/llms/vertex_ai/gemini/vertex_and_google_ai_studio_gemini.py#L362)
Added an additional non-OpenAI standard "disable" value for non-reasoning Gemini requests.
**Mapping**
| reasoning_effort | thinking |
| ---------------- | -------- |
| "disable" | "budget_tokens": 0 |
| "low" | "budget_tokens": 1024 |
| "medium" | "budget_tokens": 2048 |
| "high" | "budget_tokens": 4096 |

3
docs/my-website/docs/providers/vertex.md
View file

@ -662,10 +662,13 @@ print(resp)
LiteLLM translates OpenAI's `reasoning_effort` to Gemini's `thinking` parameter. [Code](https://github.com/BerriAI/litellm/blob/620664921902d7a9bfb29897a7b27c1a7ef4ddfb/litellm/llms/vertex_ai/gemini/vertex_and_google_ai_studio_gemini.py#L362)
Added an additional non-OpenAI standard "disable" value for non-reasoning Gemini requests.
**Mapping**
| reasoning_effort | thinking |
| ---------------- | -------- |
| "disable" | "budget_tokens": 0 |
| "low" | "budget_tokens": 1024 |
| "medium" | "budget_tokens": 2048 |
| "high" | "budget_tokens": 4096 |

3
litellm/constants.py
View file

@ -36,6 +36,9 @@ SINGLE_DEPLOYMENT_TRAFFIC_FAILURE_THRESHOLD = int(
os.getenv("SINGLE_DEPLOYMENT_TRAFFIC_FAILURE_THRESHOLD", 1000)
) # Minimum number of requests to consider "reasonable traffic". Used for single-deployment cooldown logic.
DEFAULT_REASONING_EFFORT_DISABLE_THINKING_BUDGET = int(
os.getenv("DEFAULT_REASONING_EFFORT_DISABLE_THINKING_BUDGET", 0)
)
DEFAULT_REASONING_EFFORT_LOW_THINKING_BUDGET = int(
os.getenv("DEFAULT_REASONING_EFFORT_LOW_THINKING_BUDGET", 1024)
)

6
litellm/llms/vertex_ai/gemini/vertex_and_google_ai_studio_gemini.py
View file

@ -26,6 +26,7 @@ import litellm.litellm_core_utils
import litellm.litellm_core_utils.litellm_logging
from litellm import verbose_logger
from litellm.constants import (
DEFAULT_REASONING_EFFORT_DISABLE_THINKING_BUDGET,
DEFAULT_REASONING_EFFORT_HIGH_THINKING_BUDGET,
DEFAULT_REASONING_EFFORT_LOW_THINKING_BUDGET,
DEFAULT_REASONING_EFFORT_MEDIUM_THINKING_BUDGET,
@ -428,6 +429,11 @@ class VertexGeminiConfig(VertexAIBaseConfig, BaseConfig):
"thinkingBudget": DEFAULT_REASONING_EFFORT_HIGH_THINKING_BUDGET,
"includeThoughts": True,
}
elif reasoning_effort == "disable":
return {
"thinkingBudget": DEFAULT_REASONING_EFFORT_DISABLE_THINKING_BUDGET,
"includeThoughts": False,
}
else:
raise ValueError(f"Invalid reasoning effort: {reasoning_effort}")