SAM 3 vs Gemini 3.1 Pro+ 1 other

Compare SAM 3, Gemini 3.1 Pro, and 1 other vision model side-by-side. Test these models on Object Detection in the Playground.

Compare these vision models live

Run the same image across every model that supports a task and compare their outputs side-by-side.

Detect and compare bounding boxes across models on the same image.

Open Object Detection in the full playground
MetaSAM 3
Run to compare this model.
GoogleGemini 3.1 Pro
Run to compare this model.
OpenAIGPT-5.5
Run to compare this model.

Models in this comparison

Meta
OpenAI

Model Overviews

Released on November 19th, 2025, Segment Anything 3 (SAM 3) is a zero-shot image segmentation model that “detects, segments, and tracks objects in images and videos based on concept prompts.” This model was developed by Meta as the third model in the Segment Anything series.

Unlike its previous SAM models (Segment Anything and Segment Anything 2), you can provide SAM 3 with the prompt “shipping container” and it will generate precise segmentation masks for all shipping containers in an image. SAM 3 generates segmentation masks that correspond to the location of the objects found with a text prompt.

SAM 3 vs Gemini 3.1 Pro Comparison Table + 1 other

PropertySAM 3Gemini 3.1 ProGPT-5.5
OrganizationMetaGoogleOpenAI
Categoryclosedclosedclosed
Modalitymultimodalmultimodalmultimodal
Release DateNov 2025Feb 2026Apr 2026
Context Window1.0M1.0M
Parameters
LicenseProprietaryProprietaryProprietary
Pricing per 1M tokens
Input $/1M$2.00$5.00
Output $/1M$12.00$30.00
Vision Tasks
Object DetectionDemoDemoDemo
CaptioningDemoDemo
ClassificationDemoDemo
OCRDemoDemo
Vision Language
Visual Question AnsweringDemoDemo
Instance Segmentation
Promptable Concept SegmentationDemo
Video Object Tracking
Zero Shot Segmentation
Model Features
Foundation Vision
LLMs with Vision Capabilities
Multimodal Vision
Zero-shot Detection
Vision Evalspass/fail results · 66 prompts
Score key:≥75%40–74%<40%
Visual Understanding
Overall Score
75.76%
77.61%
Avg Response Time6.13s30.12s
Median input tokensincl. image tokens1.1K1.4K
Median output tokens11138
Est. cost / taskon this benchmark$0.0024$0.011
Defect Detection
73.3%(11/15)
86.7%(13/15)
Document Understanding
88.9%(8/9)
88.9%(8/9)
Object Counting
44.4%(4/9)
30%(3/10)
Object Understanding
92.9%(13/14)
92.9%(13/14)
Spatial Understanding
73.7%(14/19)
78.9%(15/19)
OCR
Overall Score
89.52%
81.22%
Avg Response Time3.11s5.16s
Median input tokensincl. image tokens1.1K105
Median output tokens1283
Est. cost / taskon this benchmark$0.0024$0.0030
Focused Scene OCR
94.9%(94/99)
77.8%(77/99)
Handwritten Math
90%(9/10)
40%(4/10)
License Plate Recognition
90%(27/30)
93.3%(28/30)
Text Recognition
86.7%(26/30)
83.3%(25/30)
VQA & Extraction
81.7%(49/60)
86.7%(52/60)

Output tokens (incl. reasoning) and est. cost / task are measured on this benchmark from a single low-temperature run, and shown only for models whose run covered at least 90% of prompts. Methodology