GeminiNodeV3 - ComfyUI Built-in Node Documentation

Generate text responses with Google’s Gemini models. Provide a text prompt and, optionally, one or more images, audio clips, videos, or files as multimodal context. The selected model determines which additional settings appear.

Inputs

Common Inputs

Parameter	Description	Data Type	Required	Range
`model`	The Gemini model used to generate the response. The selected model determines which additional inputs are shown.	DYNAMIC_COMBO	Yes	`"Gemini 3.8 Flash"` `"Gemini 3.7 Flash"` `"Gemini 3.5 Flash"` `"Gemini 3.1 Pro"` `"Gemini 3.1 Flash-Lite"`

Media Inputs

These growable media inputs are available for every model option.

Parameter	Description	Data Type	Required	Range
`images`	Optional image(s) to use as context for the model. Up to 16 images. Growable slot: connect images to `image_1` through `image_16`.	IMAGE	No	Up to 16 images
`audio`	Optional audio clip to use as context for the model. Growable slot: `audio_1`.	AUDIO	No	1 audio clip
`video`	Optional video clip to use as context for the model. Growable slot: `video_1`.	VIDEO	No	1 video clip

Gemini 3.8 Flash Inputs

Parameter	Description	Data Type	Required	Range
`prompt`	Text input to the model. Include detailed instructions, questions, or context. Must not be empty.	STRING	Yes	Multiline text (must contain at least one non-whitespace character)
`video_processing`	How the model reads the attached video. `static` samples frames at a fixed rate and sends all of them as context; `agentic` lets the model navigate the timeline itself and load only the frames, audio or transcript it needs, which costs far fewer input tokens on long videos.	COMBO	Yes	`"static"` `"agentic"` (default: `"static"`)
`files`	Optional file(s) to use as context for the model. Accepts inputs from the Gemini Input Files node.	GEMINI_INPUT_FILES	No	N/A
`thinking_level`	How hard the model reasons internally before answering. HIGH improves quality on difficult tasks but costs more (thinking) tokens and is slower.	COMBO	Yes	`"LOW"` `"MEDIUM"` `"HIGH"` (default: `"MEDIUM"`)
`max_output_tokens`	Maximum tokens to generate, including the model’s internal thinking. With thinking_level HIGH, a low value can leave no room for the answer; raise this if responses come back empty or truncated. The model stops early when finished, so a higher cap costs nothing extra for short replies.	INT	Yes	16-65536 (default: 32768)
`seed`	Seed for sampling. Set to 0 for a random seed. Deterministic output isn’t guaranteed.	INT	Yes	0-2147483647 (default: 42)
`system_prompt`	Foundational instructions that dictate the model’s behavior.	STRING	Yes	Multiline text (default: empty)

Gemini 3.7 Flash Inputs

Parameter	Description	Data Type	Required	Range
`prompt`	Text input to the model. Include detailed instructions, questions, or context. Must not be empty.	STRING	Yes	Multiline text (must contain at least one non-whitespace character)
`files`	Optional file(s) to use as context for the model. Accepts inputs from the Gemini Input Files node.	GEMINI_INPUT_FILES	No	N/A
`thinking_level`	How hard the model reasons internally before answering. HIGH improves quality on difficult tasks but costs more (thinking) tokens and is slower.	COMBO	Yes	`"LOW"` `"MEDIUM"` `"HIGH"` (default: `"MEDIUM"`)
`temperature`	Controls randomness. Lower is more focused/deterministic, higher is more creative.	FLOAT	Yes	0.0-2.0 (default: 1.0)
`top_p`	Nucleus sampling: sample from the smallest token set whose cumulative probability reaches top_p.	FLOAT	Yes	0.0-1.0 (default: 0.95)
`max_output_tokens`	Maximum tokens to generate, including the model’s internal thinking. With thinking_level HIGH, a low value can leave no room for the answer; raise this if responses come back empty or truncated. The model stops early when finished, so a higher cap costs nothing extra for short replies.	INT	Yes	16-65536 (default: 32768)
`seed`	Seed for sampling. Set to 0 for a random seed. Deterministic output isn’t guaranteed.	INT	Yes	0-2147483647 (default: 42)
`system_prompt`	Foundational instructions that dictate the model’s behavior.	STRING	Yes	Multiline text (default: empty)

Gemini 3.5 Flash Inputs

Parameter	Description	Data Type	Required	Range
`prompt`	Text input to the model. Include detailed instructions, questions, or context. Must not be empty.	STRING	Yes	Multiline text (must contain at least one non-whitespace character)
`files`	Optional file(s) to use as context for the model. Accepts inputs from the Gemini Input Files node.	GEMINI_INPUT_FILES	No	N/A
`thinking_level`	How hard the model reasons internally before answering. HIGH improves quality on difficult tasks but costs more (thinking) tokens and is slower.	COMBO	Yes	`"MINIMAL"` `"LOW"` `"MEDIUM"` `"HIGH"` (default: `"MEDIUM"`)
`temperature`	Controls randomness. Lower is more focused/deterministic, higher is more creative.	FLOAT	Yes	0.0-2.0 (default: 1.0)
`top_p`	Nucleus sampling: sample from the smallest token set whose cumulative probability reaches top_p.	FLOAT	Yes	0.0-1.0 (default: 0.95)
`max_output_tokens`	Maximum tokens to generate, including the model’s internal thinking. With thinking_level HIGH, a low value can leave no room for the answer; raise this if responses come back empty or truncated. The model stops early when finished, so a higher cap costs nothing extra for short replies.	INT	Yes	16-65536 (default: 32768)
`seed`	Seed for sampling. Set to 0 for a random seed. Deterministic output isn’t guaranteed.	INT	Yes	0-2147483647 (default: 42)
`system_prompt`	Foundational instructions that dictate the model’s behavior.	STRING	Yes	Multiline text (default: empty)

Gemini 3.1 Pro Inputs

Parameter	Description	Data Type	Required	Range
`prompt`	Text input to the model. Include detailed instructions, questions, or context. Must not be empty.	STRING	Yes	Multiline text (must contain at least one non-whitespace character)
`files`	Optional file(s) to use as context for the model. Accepts inputs from the Gemini Input Files node.	GEMINI_INPUT_FILES	No	N/A
`thinking_level`	How hard the model reasons internally before answering. HIGH improves quality on difficult tasks but costs more (thinking) tokens and is slower.	COMBO	Yes	`"LOW"` `"HIGH"` (default: `"HIGH"`)
`temperature`	Controls randomness. Lower is more focused/deterministic, higher is more creative.	FLOAT	Yes	0.0-2.0 (default: 1.0)
`top_p`	Nucleus sampling: sample from the smallest token set whose cumulative probability reaches top_p.	FLOAT	Yes	0.0-1.0 (default: 0.95)
`max_output_tokens`	Maximum tokens to generate, including the model’s internal thinking. With thinking_level HIGH, a low value can leave no room for the answer; raise this if responses come back empty or truncated. The model stops early when finished, so a higher cap costs nothing extra for short replies.	INT	Yes	16-65536 (default: 32768)
`seed`	Seed for sampling. Set to 0 for a random seed. Deterministic output isn’t guaranteed.	INT	Yes	0-2147483647 (default: 42)
`system_prompt`	Foundational instructions that dictate the model’s behavior.	STRING	Yes	Multiline text (default: empty)

Gemini 3.1 Flash-Lite Inputs

Parameter	Description	Data Type	Required	Range
`prompt`	Text input to the model. Include detailed instructions, questions, or context. Must not be empty.	STRING	Yes	Multiline text (must contain at least one non-whitespace character)
`files`	Optional file(s) to use as context for the model. Accepts inputs from the Gemini Input Files node.	GEMINI_INPUT_FILES	No	N/A
`thinking_level`	How hard the model reasons internally before answering. HIGH improves quality on difficult tasks but costs more (thinking) tokens and is slower.	COMBO	Yes	`"LOW"` `"HIGH"` (default: `"LOW"`)
`temperature`	Controls randomness. Lower is more focused/deterministic, higher is more creative.	FLOAT	Yes	0.0-2.0 (default: 1.0)
`top_p`	Nucleus sampling: sample from the smallest token set whose cumulative probability reaches top_p.	FLOAT	Yes	0.0-1.0 (default: 0.95)
`max_output_tokens`	Maximum tokens to generate, including the model’s internal thinking. With thinking_level HIGH, a low value can leave no room for the answer; raise this if responses come back empty or truncated. The model stops early when finished, so a higher cap costs nothing extra for short replies.	INT	Yes	16-65536 (default: 32768)
`seed`	Seed for sampling. Set to 0 for a random seed. Deterministic output isn’t guaranteed.	INT	Yes	0-2147483647 (default: 42)
`system_prompt`	Foundational instructions that dictate the model’s behavior.	STRING	Yes	Multiline text (default: empty)

Note: For Gemini 3.8 Flash, temperature and top_p are not available, and video_processing is only available for this model option. The thinking_level options and default differ by model as listed above. Note: The prompt input must not be empty. The node validates that it contains at least one non-whitespace character. Note: The node uploads up to the first 10 media items as URLs, prioritizing video, then audio, then images. Any remaining media are sent inline as base64. Total inline media is limited to 18 MB; if exceeded, the node raises an error asking you to reduce the number or size of attached media.

Outputs

Output Name	Description	Data Type
`STRING`	The text response generated by the selected Gemini model.	STRING

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): d04d1e97a9c213297899291ad30db14a9f946b07506d0377fead4e29510c5ad9

​Inputs

​Common Inputs

​Media Inputs

​Gemini 3.8 Flash Inputs

​Gemini 3.7 Flash Inputs

​Gemini 3.5 Flash Inputs

​Gemini 3.1 Pro Inputs

​Gemini 3.1 Flash-Lite Inputs

​Outputs