LOADING DATASHEET
LOADING DATASHEET
by Meta
RANKED #227 OF 337 ASSISTANTS · OVERALL #360 OF 528 RANKED · VIBE SCORE 5.5 · 32 VOICES
COPY BADGE pastes a README snippet with the rank SVG.
SAM 3 is a unified foundation model for promptable segmentation in images and videos. It can detect, segment, and track objects using text or visual prompts such as points, boxes, and masks.
10 mentions
2 mentions
26 weeks · 25 voices
AGING WELL
The crowd is warmer now than it was at the start.
Open weights. You can use a hosted service, or download it and run it yourself, free.
Open weights: download Sam 3 once and it's yours. No subscription, no rate limits, works offline.
840M parameters, full precision PyTorch weights are about 3.4 GB
NVIDIA GeForce RTX 3060 12GB
12 GB VRAM
~$260 USEDFP16 image segmentation and short video tracking, ~0.08 s/image
NVIDIA GeForce RTX 4070 SUPER 12GB
12 GB VRAM
~$590 NEWFP16 batched promptable concept segmentation and smooth video inference, ~0.04 s/image
NVIDIA GeForce RTX 4090 24GB
24 GB VRAM
~$1,750 NEWFull-precision high-resolution batch segmentation and long multi-object video tracking, ~0.02 s/image
The three cards above are researched picks. Search the machine you actually have — we only say yes if it should reply at a usable speed, not just load the weights and crawl.
HARDWARE PICKS & PRICES ARE RESEARCHED FROM THE LIVE WEB AND REFRESHED AUTOMATICALLY. TREAT THEM AS BALLPARK, NOT GOSPEL.
Everyday jobs on each of those rigs: how long Sam 3 takes, and how good the crowd says it is at that kind of work.
| THE JOB | BARE MINIMUMNVIDIA GeForce RTX 3060 12GB | SWEET SPOTNVIDIA GeForce RTX 4070 SUPER 12GB | FULL POWERNVIDIA GeForce RTX 4090 24GB |
|---|---|---|---|
| Write an emaila paragraph or two, drafted from a one-line brief | NO DATA | NO DATA | NO DATA |
| Summarize a documenta long report boiled down to the points that matter | —★★★★★ | —★★★★★ | —★★★★★ |
| Build a websitea small landing page, markup and styles together | —★★★★★ | —★★★★★ | —★★★★★ |
| Build a backendan API with routes, storage and tests | —★★★★★ | —★★★★★ | —★★★★★ |
| Build a gamea playable browser game in one file | —★★★★★ | —★★★★★ | —★★★★★ |
STARS ARE THE COMMUNITY SCORE FOR THAT KIND OF WORK, DOCKED FOR HOW SQUASHED THE WEIGHTS GET AT EACH TIER. A DASH MEANS NO SPEED WAS RESEARCHED FOR THAT RIG. TIMES ARE COMPUTED FROM RESEARCHED SPEEDS. BALLPARK, NOT GOSPEL.
Not enough discussion about Sam 3 yet to call it. We found 32 posts but people did not say much either way.
24 POSTS · 8 COMMENTS · REDDIT · X · HACKER NEWS · DEV.TO · GITHUB
Other models the crowd has fully reviewed, starting with text models like this one.
Brand NEW Meta SAM3 - now for Comfy-UI !
SAM3 added to Comfyui-Angelo (sampler/inpainter/refiner)
Vision banana!!!!
SCAIL-2/SAM 3 Tracking Help
Show HN: MultiMatte, a Promptable Image Background Removal Model
[X-post] AMA with the Meta researchers behind SAM 3 + SAM 3D + SAM Audio
Use SAM3 to Segment Subjects for Precise Image Editing When Your Model Doesn’t Support Inpainting (Demo Included)
AMA with the Meta researchers behind SAM 3 + SAM 3D + SAM Audio
Comfyui workflow to generate expressions
Sam 3 for Wan Animate
Best identity swapper workflow or model/LoRA?
Gemini 3.7 Flash Extended cant stop hallucinating on first prompt