google/gemini-2.5-flash 🔢🖼️📝✓ → 📝

⭐ Official ▶️ 778.6K runs 📅 Oct 2025 ⚙️ Cog 0.16.9
code-generation image-analysis image-captioning image-to-text question-answering text-generation text-translation video-to-text visual-understanding

About

Google’s hybrid “thinking” AI model optimized for speed and cost-efficiency

Example Output

Prompt:

"describe this image in detail"

Output

This stunning image depicts a solitary samurai warrior in a moment of deep contemplation amidst a breathtaking autumn mountain landscape, rendered in a style reminiscent of classical painting with a touch of modern digital artistry.

In the foreground, slightly to the left of center, a samurai is seated on a rugged, dark gray and brown rocky outcrop. He is in a contemplative or meditative posture, with his head bowed and gaze directed downwards. His hands are clasped or resting gently in his lap.

The samurai's armor is remarkably ornate and colorful. He wears a regal purple kabuto (helmet) with golden, horn-like kuwagata crests on the front, adorned with intricate golden details. His face is obscured by a dark blue or black mask (mempo), with subtle red accents around what appears to be the mouth or chin guard. The body armor consists of white or cream-colored sections with delicate, intricate purple patterns, possibly depicting floral motifs, clouds, or traditional Japanese designs. These patterned sections are interspersed with solid purple plates, all appearing to be meticulously lacquered and held together with visible lacing. A katana, sheathed in a dark scabbard, rests at his left hip.

Dominating the left and upper-middle sections of the image is a gnarled, ancient tree branch, its dark, textured bark contrasting with its vibrant foliage. The leaves are a magnificent tapestry of autumn colors: fiery reds, rich oranges, and brilliant yellows, suggesting a maple or similar deciduous tree. The individual leaves are highly detailed, showing veins and variations in color, and are clustered densely on the branches, some smaller twigs extending with nascent buds or berries.

The background unfolds into a majestic mountain range. Several layers of peaks fade into the distance, rendered with atmospheric perspective, appearing progressively lighter blue and hazier. The most prominent feature in the background is a towering, sharp, snow-capped mountain peak on the far right, its pristine white snow standing out against the pale, almost white or light blue sky.

The overall color palette is a striking blend of vibrant autumn hues (reds, oranges, yellows) in the foliage, deep purples, whites, and dark blues of the samurai's armor, and the cool blues, grays, and whites of the distant mountains and sky. The lighting is soft and diffused, creating a serene and introspective mood. The image evokes a sense of peace, resilience, and the beauty of nature, with the samurai serving as a stoic, reflective presence within the grand landscape.

Performance Metrics

10.63s Prediction Time
10.65s Total Time
All Input Parameters
{
  "top_p": 0.95,
  "images": [
    "https://replicate.delivery/pbxt/O1TXLIqsDC7pdOzMl259hqrdwkDLjxrf8Fsg2ZwYVIIkoHkm/replicate-prediction-vygd5qqab1rmc0ctb9cbr14cxw.jpg"
  ],
  "prompt": "describe this image in detail",
  "videos": [],
  "temperature": 1,
  "dynamic_thinking": false,
  "max_output_tokens": 65535
}
Input Parameters
top_p Type: numberDefault: 0.95Range: 0 - 1
Nucleus sampling parameter - the model considers the results of the tokens with top_p probability mass
images Type: arrayDefault:
Input images to send with the prompt (max 10 images, each up to 7MB)
prompt (required) Type: string
The text prompt to send to the model
videos Type: arrayDefault:
Input videos to send with the prompt (max 10 videos, each up to 45 minutes)
temperature Type: numberDefault: 1Range: 0 - 2
Sampling temperature between 0 and 2
thinking_budget Type: integerRange: 0 - 24576
Thinking budget for reasoning (0 to disable thinking, higher values allow more reasoning)
dynamic_thinking Type: booleanDefault: false
Enable dynamic thinking - the model will adjust the thinking budget based on the complexity of the request (overrides thinking_budget parameter)
max_output_tokens Type: integerDefault: 65535Range: 1 - 65535
Maximum number of tokens to generate
system_instruction Type: string
System instruction to guide the model's behavior
Output Schema

Output

Type: arrayItems Type: string

Example Execution Logs
Generating text with gemini-2.5-flash...
Processing 1 images...
Text generation completed in 10.57s
Input token count: 263
Output token count: 515
Version Details
Version ID
bfb7df9586ae4fafa00a593d8dc4868698f72cf9d695da28b8c8a70f88e876ba
Version Created
December 16, 2025
Run on Replicate →