Segment Anything Model 3
Hacker ApprovedSegmentation model in AI tools
People recommend it for fast, accurate zero-shot segmentation, streaming video support, and as a teacher model for training smaller ones. Critics say it fits specialised work such as engineering drawings and medical imaging poorly.
People recommend it for fast, accurate zero-shot segmentation, streaming video support, and as a teacher model for training smaller ones. Critics say it fits specialised work such as engineering drawings and medical imaging poorly.
57 people gave an opinion
- recommend it
- 48
- criticize it
- 9
- approval share
- 84%
Recommended in 9 of 13 months
- November 2025: recommended
- December 2025: recommended
- January 2026: recommended
- February 2026: not recommended
- March 2026: recommended
- April 2026: not recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: not recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it carries the badge
- 48 people recommend itNeeds 10 or more, met.
- 5.3 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 9 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Seen as a breakthrough in segmentation
“It's really, really good. This feels like a seminal moment for computer vision.”
Comment from 19 Nov 2025 in “Meta Segment Anything Model 3” “Segment Anything literaly leapfroged progress on image segmentation.”
Comment from 1 Dec 2025 in “It’s been a very hard year” “Segment Anything 2 was a breakthrough in image segmentation”
Comment from 26 Nov 2025 in “OpenAI needs to raise at least $207B by 2030” - Comment from 20 Nov 2025 in “Meta Segment Anything Model 3”
Accurate zero-shot segmentation with text prompts
“Nowadays that's no problem anymore as you can just use models like meta's segment anything that do zero-shot bounding boxes”
Comment from 10 Nov 2025 in “The overengineered solution to my pigeon problem (2022)” “this model is extremely good - the "zero-shot" text prompted detection is a huge step ahead of what we've seen before”
Comment from 19 Nov 2025 in “Meta Segment Anything Model 3” “it does pretty well. At least better than other available models”
Comment from 19 Nov 2025 in “Meta Segment Anything Model 3” Native streaming video support
“SAM3 has built-in streaming support which is very exciting.”
Comment from 19 Nov 2025 in “Meta Segment Anything Model 3” “The native support for streaming in SAM3 is awesome.”
Comment from 19 Nov 2025 in “Meta Segment Anything Model 3” Teacher model for distilling smaller ones
“SAM3 is probably a great model to distill from when training smaller segmentation models”
Comment from 19 Nov 2025 in “Meta Segment Anything Model 3” “it's so flexible and accurate it's an amazing teacher for lean CNNs”
Comment from 20 Nov 2025 in “Meta Segment Anything Model 3”
What people warn about
Each mark is one comment
Weak on specialised domains
“It's great with a pony on the beach but really isn't made for more rote industrial type applications.”
Comment from 19 Nov 2025 in “Meta Segment Anything Model 3” “These models struggle, understandably so, with medical imaging.”
Comment from 20 Nov 2025 in “Meta Segment Anything Model 3” “SAM3 works bad on engineering drawings while SAM2 kinda works”
Comment from 25 Nov 2025 in “Meta Segment Anything Model 3” Doesn't fit some specific use cases
“It is an amazing model but it doesn't work for my use case”
Comment from 20 Nov 2025 in “Meta Segment Anything Model 3” “This fantastic sam model is maybe not the right fit for our application.”
Comment from 20 Nov 2025 in “Meta Segment Anything Model 3”
What people said
40 of the 67 comments that mention it, newest first. Each links to the original comment.
“SAM on its own can't really decide what you need to segment”
“We already got a lot better vision ML algorithm because of this (tx to suckerbergs dino and segment anything)”
“Considering all of the great research that has come from his labs”
“SAM2 supports coordinate based masks (user taps or clicks somewhere in an image)”
“They can however be extremely useful for curating training data.”
“What I'm actually excited about in AI are smaller specialised things like whisper, CLIP, SAM, etc.”
“Those SAM models they produce deliver more value than a hypothetical competitive llama5 model coming out tomorrow.”
“Their VGGT, Dinov3, and segment anything models are pretty impressive.”
“Segment Anything literaly leapfroged progress on image segmentation.”
“Segment Anything 2 was a breakthrough in image segmentation”
“This would be convenient for post-production and editing of video”
“SAM3 works bad on engineering drawings while SAM2 kinda works”
“Segment anything however was able to segment all 5 dog legs when prompted to.”
“But open sourcing the weights help move the field in general.”
“it has been great for experimenting without the thinking overhead of paying too much for inference”
“it's so flexible and accurate it's an amazing teacher for lean CNNs”
“If they really deliver a model that can track and describe existing images / videos well that would be a huge breakthrough.”
“I'm impressed by the ability of this model to create a image encoding that is independent of the prompt”
“I like the result but let's not pretend it's for gracious intent.”
“Revolutionized image segmentation overnight”
“These models struggle, understandably so, with medical imaging.”
“This fantastic sam model is maybe not the right fit for our application.”
“actions like this are much appreciated and benefit everyone”
“It is an amazing model but it doesn't work for my use case”
“it seems to do well, but it just gives me the image, I see no way to export or use this image”
“It's great with a pony on the beach but really isn't made for more rote industrial type applications.”
“With the default SAM2 and loading images from the da, we could only process videos with 10^2 - 10^3 frames before running out of memory.”
“The native support for streaming in SAM3 is awesome.”
“SAM3 is probably a great model to distill from when training smaller segmentation models”
“seems like it would work great for game asset generation regarding 3D”
“SAM3 seems to less precisely trace the images — it'll discard kids drawing out the lines a bit”
“I'm bullish SAM 3 can automate at least 90% of the work.”
“I couldn't download it. Model appears to be comparable to Sparc3D, Huyunan, etc but w/o download, who can say? It is much faster though.”
“this model is extremely good - the "zero-shot" text prompted detection is a huge step ahead of what we've seen before”
“This is an incredible model. But once again, we find an announcement for a new AI model with highly misleading graphs.”
“SAM3 has built-in streaming support which is very exciting.”
“it does pretty well. At least better than other available models”
“It's really, really good. This feels like a seminal moment for computer vision.”
“SAM is great, what value does it bring to META business?”
“Nowadays that's no problem anymore as you can just use models like meta's segment anything that do zero-shot bounding boxes”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.