| we've been testing astra across domains and are finding some surprises and specific areas it is reliable
exciting zero… |
@roboflow |
Company |
Original |
2026-09-10 |
43,232 |
449 |
24 |
17 |
2 |
401 |
| the Driveline team is pushing the frontier of what's possible with AI + sports
if you're interested in how AI is shapi… |
@roboflow |
Company |
Quote |
2026-09-08 |
7,238 |
48 |
5 |
3 |
0 |
12 |
| you can label data in roboflow with astra
millions of vision products are now unblocked by labeling time/cost going al… |
@roboflow |
Company |
Quote |
2026-09-07 |
38,141 |
478 |
38 |
7 |
3 |
357 |
| GPT-6 Astra is now the top model for vision tasks
sure 3-10x more expensive but 50%+ faster
extremely capable model, … |
@roboflow |
Company |
Quote |
2026-09-06 |
7,496 |
119 |
9 |
5 |
2 |
22 |
| new Gemini releases aren't moving much on the vision leaderboard in terms of accuracy scores, Gemini 3.5 Flash is still… |
@roboflow |
Company |
Quote |
2026-09-03 |
3,045 |
32 |
0 |
0 |
0 |
12 |
| Gemini 3.8 Flash and Fable 5.1 vision evals are live, new top 10 ranking order
Google still dominating
Fable improvin… |
@roboflow |
Company |
Quote |
2026-09-02 |
3,104 |
32 |
1 |
3 |
1 |
7 |
| a great way to ensure your microduck gets plenty of exercise |
@roboflow |
Company |
Quote |
2026-09-01 |
1,637 |
12 |
1 |
0 |
0 |
3 |
| it's really hard to understand and estimate costs for vision tasks with frontier models
adding features to make it as … |
@roboflow |
Company |
Quote |
2026-08-25 |
1,389 |
12 |
0 |
1 |
0 |
0 |
| muse spark 1.2 is a top model for real world vision use cases as well
meta models and google models dominate the top o… |
@roboflow |
Company |
Quote |
2026-08-20 |
2,239 |
17 |
1 |
2 |
0 |
3 |
| We are hosting the Visual Intelligence Summit on October 22 in SF.
The physical world is AI's biggest opportunity and … |
@roboflow |
Company |
Original |
2026-08-19 |
52,209 |
55 |
15 |
7 |
9 |
14 |
| RT @IAStartupWeek: Announcing Joseph Nelson, CEO & Co-Founder of Roboflow as our Keynote Speaker for Iowa Startup W… |
@roboflow |
Company |
Repost |
2026-08-19 |
102 |
0 |
1 |
0 |
0 |
0 |
| full evals are now live for grok 4.6
for sure a model to consider depending on your use case, cost, and latency
you c… |
@roboflow |
Company |
Quote |
2026-08-13 |
1,290 |
10 |
0 |
1 |
0 |
3 |
| had the chance to start evaluating grok 4.6 earlier this week, our vision eval is tough real world data only
4.6 is ge… |
@roboflow |
Company |
Quote |
2026-08-12 |
7,184 |
30 |
1 |
2 |
3 |
6 |
| tracking is core to any video analytics use cases. you need the added accuracy for understanding complex scenes
our tr… |
@roboflow |
Company |
Quote |
2026-08-12 |
6,103 |
101 |
8 |
0 |
0 |
87 |
| our deep dive on qwen3.8-max
lots of examples and breakdowns by each category we eval, it's a top choice for vision ta… |
@roboflow |
Company |
Quote |
2026-08-06 |
1,958 |
20 |
1 |
0 |
0 |
7 |
| frontier models are handling more and more complex tasks without fine-tuning
models like qwen-3.8 are shifting the ROI… |
@roboflow |
Company |
Quote |
2026-08-05 |
1,147 |
9 |
1 |
4 |
0 |
4 |
| vision evals leaderboard looks totally different from 30 days ago
3 new frontier models from 3 different labs now beat… |
@roboflow |
Company |
Quote |
2026-08-04 |
2,278 |
28 |
2 |
3 |
0 |
5 |
| big breakthrough: qwen3.8-max is now the top of our object detection leaderboard
if the latency works for your use cas… |
@roboflow |
Company |
Quote |
2026-08-03 |
3,539 |
20 |
1 |
2 |
1 |
9 |
| pretty clear at this point what model you should use to automate labeling vision data |
@roboflow |
Company |
Quote |
2026-07-30 |
1,126 |
6 |
1 |
0 |
0 |
2 |
| just expanded our image reasoning evals and gemini 3.5 flash is still at the top
best frontier model for visual tasks … |
@roboflow |
Company |
Quote |
2026-07-29 |
1,741 |
15 |
1 |
3 |
0 |
2 |
| gemini 3.6 flash breakdown with example images
key finding: 3.5 Flash-Lite beats Fable 5 while being ~25x cheaper
gem… |
@roboflow |
Company |
Quote |
2026-07-22 |
1,632 |
9 |
1 |
2 |
0 |
3 |
| we benchmarked inkling on vision tasks, it's coming in at the bottom of the leaderboard. probably expected given they a… |
@roboflow |
Company |
Quote |
2026-07-21 |
1,968 |
10 |
0 |
3 |
0 |
3 |
| very good thread with sample data showing Fable 5 performance against other models for vision tasks
important to see t… |
@roboflow |
Company |
Quote |
2026-07-02 |
3,085 |
14 |
0 |
2 |
0 |
5 |
| Fable 5 is back and tied as a top vision model, Sonnet 5 is outside the top 10
Fable 5 is ~3x slower and ~10x more exp… |
@roboflow |
Company |
Quote |
2026-07-02 |
6,822 |
30 |
4 |
3 |
1 |
4 |
| every legacy vision product/system/company can probably be built with open source tools and models at this point
and t… |
@roboflow |
Company |
Quote |
2026-07-01 |
6,474 |
59 |
4 |
1 |
0 |
22 |
| We added video understanding and Gemini 3.5 Flash wins there as well
Seems like the all around best frontier vision mo… |
@roboflow |
Company |
Quote |
2026-06-30 |
2,288 |
19 |
6 |
1 |
0 |
5 |
| rf-detr + sports never misses |
@roboflow |
Company |
Quote |
2026-06-24 |
9,796 |
96 |
4 |
1 |
0 |
56 |
| threads like this where ai projects snowball into the craziest demos of all time is exactly what we are here for |
@roboflow |
Company |
Quote |
2026-06-20 |
6,218 |
23 |
1 |
1 |
0 |
11 |
| a few things that make this exciting
- 7 images to get a custom model means nobody is constrained by data or compute t… |
@roboflow |
Company |
Quote |
2026-06-16 |
11,932 |
134 |
16 |
2 |
1 |
77 |
| any requests for new apps we should build open source so you can build a startup on top?
or maybe, what companies do y… |
@roboflow |
Company |
Quote |
2026-06-15 |
2,256 |
6 |
2 |
3 |
0 |
7 |
| You can now build vision applications with Fable 5 in Roboflow
Fable 5 is the best Anthropic model for visual understa… |
@roboflow |
Company |
Quote |
2026-06-11 |
10,812 |
43 |
2 |
3 |
1 |
15 |
| supervision is #3 trending repo today and crossed 40k stars a few days ago, we built it to make computer vision dead si… |
@roboflow |
Company |
Quote |
2026-06-09 |
3,264 |
37 |
5 |
0 |
0 |
25 |
| No data, no problem
introducing agentic synthetic data generation with Cosmos 3
share a few examples, generate more d… |
@roboflow |
Company |
Quote |
2026-06-01 |
39,508 |
171 |
27 |
7 |
4 |
107 |
| rf-detr helping make big gains for small object tracking
detection rate from 47% to 86% |
@roboflow |
Company |
Quote |
2026-05-28 |
2,931 |
29 |
2 |
2 |
0 |
20 |
| You can now use models on OpenRouter directly in Roboflow
New models are dropping weekly and it's important to build y… |
@roboflow |
Company |
Quote |
2026-05-26 |
2,778 |
16 |
2 |
2 |
0 |
7 |
| we saw the same thing with our vision leaderboard
still testing and wrapping our heads around Gemini 3.5 Flash but it … |
@roboflow |
Company |
Quote |
2026-05-22 |
19,518 |
144 |
12 |
5 |
3 |
52 |
| well well well look what the 🐐 delivered
cool to see rf-detr generalize well across sports and camera angles
pipeline… |
@roboflow |
Company |
Quote |
2026-05-20 |
15,394 |
106 |
6 |
3 |
0 |
77 |
| here's your sign that AI for vision is very early and you should be building things |
@roboflow |
Company |
Quote |
2026-05-05 |
2,911 |
17 |
0 |
0 |
0 |
2 |
| lots of debate about Claude Opus 3.7 and GPT-5.5 for coding
if you're curious which is better for vision tasks, we ran… |
@roboflow |
Company |
Quote |
2026-05-04 |
4,689 |
25 |
4 |
3 |
0 |
3 |
| RF-DETR with the Apache 2.0 license is a really good deal |
@roboflow |
Company |
Quote |
2026-04-28 |
5,878 |
81 |
2 |
2 |
1 |
14 |
| the level of fidelity and craft going into understanding sports at a deeper level is getting crazy |
@roboflow |
Company |
Quote |
2026-04-27 |
8,480 |
92 |
5 |
1 |
0 |
57 |
| Roboflow MCP is starting to take shape
thanks to the early folks using and giving feedback. Still limited scope but go… |
@roboflow |
Company |
Quote |
2026-04-24 |
4,314 |
46 |
7 |
3 |
2 |
23 |
| great step by step view into creating a vision application by one of the most creative builders in AI |
@roboflow |
Company |
Quote |
2026-04-24 |
1,990 |
17 |
1 |
1 |
0 |
6 |
| Better tracking leads to more accurate counting.
We've been benchmarking and optimizing trackers like SORT, OC-SORT, a… |
@roboflow |
Company |
Quote |
2026-04-21 |
4,006 |
51 |
1 |
1 |
0 |
28 |
| rolling out neural architecture search for RF-DETR to accounts over the next 2 weeks, starting with paid accounts
....… |
@roboflow |
Company |
Quote |
2026-04-20 |
3,646 |
33 |
2 |
8 |
0 |
11 |
| cool to see companies building visual agents using Roboflow Universe models
there are over 500k fine-tuned models on u… |
@roboflow |
Company |
Quote |
2026-04-17 |
1,963 |
26 |
3 |
1 |
0 |
9 |
| With Claude Opus 4.7, Anthropic has it's first model in the top 10 of the vision leaderboard.
We track 57 VLMs across… |
@roboflow |
Company |
Quote |
2026-04-16 |
1,263 |
14 |
2 |
1 |
2 |
6 |
| When Dr Satya Mallick is happy, we are happy
Our paper on Neural Architecture Search (NAS) was accepted to ICLR 2026 a… |
@roboflow |
Company |
Quote |
2026-04-15 |
2,287 |
23 |
2 |
1 |
0 |
4 |
| Added Gemma 4 (31B and 26B/A4B), Qwen 3.6 Plus, LFM 2.5 to the vision leaderboard
Gemma 4 is Google's best open model,… |
@roboflow |
Company |
Quote |
2026-04-10 |
5,515 |
52 |
2 |
2 |
0 |
28 |
| here's what you can build for $0.00 with 3 open source models
token cost breakdown below and the company getting rich… |
@roboflow |
Company |
Quote |
2026-04-06 |
24,320 |
362 |
39 |
2 |
1 |
321 |
| this is why we built vision evals and vision arena
our team needed tools to understand new model capabilities and a fa… |
@roboflow |
Company |
Quote |
2026-04-01 |
1,804 |
13 |
2 |
2 |
0 |
3 |
| latest SAM 3 release hits 30+ FPS on an H100
one of the best zero shot detection models is now available for real-time… |
@roboflow |
Company |
Quote |
2026-03-30 |
10,301 |
109 |
8 |
4 |
0 |
61 |
| no monies no problem
RF-DETR is SOTA and apache 2.0
we built it for the builders, no tricky licensing |
@roboflow |
Company |
Quote |
2026-03-27 |
6,941 |
78 |
10 |
3 |
1 |
29 |
| seeing more and more vision models being deployed on mobile
repos like React Native ExecuTorch are making it easier fo… |
@roboflow |
Company |
Quote |
2026-03-26 |
2,495 |
29 |
2 |
1 |
1 |
15 |
| fun tutorial on building a vision powered real-time content moderation application
...also a sneak preview of an unrel… |
@roboflow |
Company |
Quote |
2026-03-25 |
2,221 |
26 |
5 |
1 |
1 |
14 |
| rf-detr segemtnation running 30+ fps
gm to the vision builders |
@roboflow |
Company |
Quote |
2026-03-24 |
8,342 |
102 |
10 |
3 |
0 |
65 |
| added Cursor cli with Composer 2 to the eval based on community feedback
#2 coding agent for vision tasks
updated blo… |
@roboflow |
Company |
Quote |
2026-03-20 |
2,374 |
31 |
2 |
1 |
1 |
9 |
| claude code is the best coding agent for building vision applications right now based on our initial set of evals
just… |
@roboflow |
Company |
Quote |
2026-03-17 |
7,072 |
63 |
8 |
4 |
1 |
26 |
| new trackers-2.3.0 is out and supports OCR-SORT
see how OC-SORT performs against SORT and ByteTrack across MOT17, Spor… |
@roboflow |
Company |
Quote |
2026-03-16 |
8,625 |
134 |
5 |
3 |
0 |
68 |
| DGX Spark provides a lot of horsepower for really high quality detection + tracking with RF-DETR
@skalskip92 is workin… |
@roboflow |
Company |
Quote |
2026-03-12 |
1,961 |
20 |
0 |
1 |
0 |
12 |
| RF-DETR is open source and powers real-time apps and analytics on your laptop
you don't need access to big GPUs to bui… |
@roboflow |
Company |
Quote |
2026-03-11 |
2,883 |
40 |
2 |
1 |
0 |
19 |
| AA delivers another banger heading into the weekend
who needs a midi when you have scissors, spoon, bowl, and RF-DETR |
@roboflow |
Company |
Quote |
2026-03-06 |
1,764 |
20 |
0 |
1 |
0 |
6 |
| just added 10 new models to the vision leaderboard, including all the small Qwen-3.5 models
new models: GPT-5.4, Grok-… |
@roboflow |
Company |
Quote |
2026-03-06 |
11,431 |
79 |
8 |
8 |
1 |
33 |
| joel gets it |
@roboflow |
Company |
Quote |
2026-03-03 |
5,054 |
34 |
1 |
2 |
0 |
14 |
| they built this in a week!
cool mix of technologies, vision + AR is a fun combo
would love to see simulators for spor… |
@roboflow |
Company |
Quote |
2026-03-03 |
2,724 |
27 |
3 |
0 |
0 |
9 |
| Qwen 3.5 is now #5 on the vision leaderboard
and it is super fast compared to the rest, 10x+ faster
Seems like with a… |
@roboflow |
Company |
Quote |
2026-03-02 |
13,735 |
108 |
8 |
6 |
1 |
36 |
| tracking tiny fast moving objects used to be hard
now it just takes
- RF-DETR
- 45 annotated images
- Claude
vision i… |
@roboflow |
Company |
Quote |
2026-02-27 |
5,180 |
69 |
3 |
2 |
0 |
41 |
| getting AI performant enough on mobile was part of the vision for RF-DETR
so many use cases to be built, greenfield op… |
@roboflow |
Company |
Quote |
2026-02-26 |
3,781 |
44 |
1 |
1 |
0 |
22 |
| Gemini 3.1 with Tools (released yesterday) set the new high score for the vision leaderboard
Google and OpenAI are bui… |
@roboflow |
Company |
Quote |
2026-02-20 |
10,370 |
26 |
5 |
2 |
2 |
8 |
| trackers livestream starting now
live coding, Q&A, latest models
join us! |
@roboflow |
Company |
Quote |
2026-02-19 |
1,742 |
28 |
0 |
0 |
0 |
10 |
| every time we release something, a new use case pops up within 24hrs and it's way cooler than our demo
gotta love open… |
@roboflow |
Company |
Quote |
2026-02-19 |
4,132 |
52 |
1 |
1 |
0 |
18 |
| we are trying to make tracking plug and play, not easy but we are working on it
big step today to make that happen wit… |
@roboflow |
Company |
Quote |
2026-02-18 |
3,038 |
50 |
2 |
5 |
0 |
19 |
| making projects open sources leads to cool stuff like this happening in just a weekend
taking the basketball ai tutori… |
@roboflow |
Company |
Quote |
2026-02-18 |
8,883 |
81 |
8 |
2 |
2 |
47 |
| vision models running in the browser in real-time are going to add so much new intelligence to big industries
sports, … |
@roboflow |
Company |
Quote |
2026-02-17 |
16,099 |
120 |
11 |
1 |
0 |
82 |
| there's so much to learn when it comes to vision
we are going deep on tracking and finding some cool stuff
tons of sm… |
@roboflow |
Company |
Quote |
2026-02-17 |
6,199 |
73 |
5 |
1 |
0 |
23 |
| we kept running into the same thing every time a new VLM came out so we built an open source community benchmark
easy … |
@roboflow |
Company |
Quote |
2026-02-16 |
6,103 |
38 |
4 |
5 |
3 |
28 |
| ummm idk about that one mate, just doing it a bit different in the west
https://t.co/6rX9FxyfT9 https://t.co/b1GwhOpnIj |
@roboflow |
Company |
Quote |
2026-02-16 |
2,398 |
15 |
0 |
0 |
0 |
3 |
| breakdown of RF-DETR and why transformers are the future of the vision frontier
worth watching if you've only used YOL… |
@roboflow |
Company |
Quote |
2026-02-16 |
2,879 |
36 |
0 |
4 |
0 |
21 |
| lots of work on tracking being done, more releases soon
RF-DETR + SAHI + ByteTrack makes detecting small objects way m… |
@roboflow |
Company |
Quote |
2026-02-13 |
15,280 |
207 |
22 |
4 |
0 |
140 |
| W for the vision community to have folks building cool stuff like this in public
if you're not following @mfranz_on , … |
@roboflow |
Company |
Quote |
2026-02-13 |
1,526 |
16 |
0 |
2 |
0 |
10 |
| In the 2020s many were certain YOLO was going to be the best architecture for vision |
@roboflow |
Company |
Quote |
2026-02-13 |
1,645 |
17 |
2 |
2 |
0 |
0 |
| Most "AI Trends" are just people guessing. We decided to look at the data instead.
We analyzed 200,000 real world visi… |
@roboflow |
Company |
Original |
2026-02-12 |
66,676 |
43 |
11 |
2 |
4 |
25 |
| great breakdown of how we achieved SOTA performance with RF-DETR
Neural architecture search is coming to the platform … |
@roboflow |
Company |
Quote |
2026-02-10 |
1,305 |
17 |
3 |
3 |
0 |
4 |
| vercel ai accelerator is back
$40k in roboflow credits to all participants
awesome program, highly recommend |
@roboflow |
Company |
Quote |
2026-02-09 |
1,949 |
8 |
1 |
1 |
0 |
4 |
| we are building olympics projects all week
any requests or ideas? |
@roboflow |
Company |
Quote |
2026-02-03 |
5,508 |
77 |
6 |
12 |
0 |
31 |
| Olympics + vision demos have started! Roboflow Rapid is powering this one.
We are working on some this week as well. J… |
@roboflow |
Company |
Quote |
2026-02-02 |
1,571 |
16 |
1 |
1 |
0 |
6 |
| Gemini 3 Flash with Tools officially topping the VLM leaderboard for real world vision tasks
Biggest jump in performan… |
@roboflow |
Company |
Quote |
2026-01-29 |
4,500 |
48 |
5 |
4 |
0 |
16 |
| Piotr is going live in ~30min to talk segmentation + tracking
Live coding session for building real-time segmentation … |
@roboflow |
Company |
Quote |
2026-01-29 |
1,328 |
16 |
1 |
1 |
0 |
3 |
| winter olympics start in 2 weeks
ideas for cool vision projects we should build? |
@roboflow |
Company |
Quote |
2026-01-28 |
1,771 |
12 |
0 |
12 |
0 |
7 |
| building with permissive open source licenses or ones that can get you sued
which way bud? https://t.co/VrB8eeaTn1 |
@roboflow |
Company |
Quote |
2026-01-15 |
2,143 |
14 |
2 |
1 |
0 |
5 |
| new YOLO released ~24 hrs ago, we have two Colabs and full platform support live for you today.
join Piotr in 15 minut… |
@roboflow |
Company |
Quote |
2026-01-15 |
1,563 |
14 |
1 |
2 |
0 |
6 |
| On the twelfth day of #Shipmas, Roboflow gave to me... Qwen3-VL, fine-tuned and deployed.
Qwen3-VL is now in Roboflow … |
@roboflow |
Company |
Original |
2026-01-06 |
5,089 |
88 |
8 |
6 |
1 |
22 |
| On the eleventh day of #Shipmas, Roboflow gave to me... Depth Anything v3.
Monocular depth maps are now in Roboflow In… |
@roboflow |
Company |
Original |
2026-01-05 |
2,968 |
21 |
3 |
3 |
1 |
7 |
| On the tenth day of #Shipmas, Roboflow gave to me... MJPEG camera streams.
Now even old-school IP cameras can plug str… |
@roboflow |
Company |
Original |
2026-01-03 |
13,446 |
309 |
12 |
7 |
2 |
76 |
| On the ninth day of #Shipmas, Roboflow gave to me... edge deploys made easy.
Today is Edge Day:
- JetPack 6.2 support
… |
@roboflow |
Company |
Original |
2026-01-02 |
3,151 |
23 |
1 |
1 |
2 |
8 |
| On the eighth day of #Shipmas, Roboflow gave to me… focus checks on screen.
Our updated Camera Focus Block adds a Tene… |
@roboflow |
Company |
Original |
2026-01-01 |
5,192 |
62 |
3 |
2 |
1 |
30 |
| On the seventh day of #Shipmas, Roboflow gave to me... s̶e̶v̶e̶n̶ ̶s̶w̶a̶n̶s̶ the best way to try, compare, and build with multimo… |
@roboflow |
Company |
Original |
2025-12-31 |
2,003 |
17 |
0 |
2 |
1 |
4 |
| On the sixth day of #Shipmas, Roboflow gave to me... six pics a-texting.
We've updated our Twilio Workflows Block to s… |
@roboflow |
Company |
Original |
2025-12-30 |
1,461 |
4 |
0 |
1 |
1 |
1 |
| On the fifth day of #Shipmas, Roboflow gave to me... five golden clicks.
Roboflow Rapid now has full annotation contro… |
@roboflow |
Company |
Original |
2025-12-29 |
2,510 |
18 |
0 |
1 |
2 |
6 |
| On the fourth day of #Shipmas, Roboflow gave to me more Trending finds.
Introducing Trending: a daily-refreshed homep… |
@roboflow |
Company |
Original |
2025-12-28 |
2,747 |
33 |
3 |
1 |
1 |
4 |
| On the third day of #Shipmas, Roboflow gave to me: motion detection in Workflows.
Our new a motion detection feature e… |
@roboflow |
Company |
Original |
2025-12-27 |
5,894 |
166 |
5 |
6 |
2 |
10 |
| On the second day of #Shipmas, Roboflow gave to me: new advanced features.
Roboflow is built for everyone, but it's al… |
@roboflow |
Company |
Original |
2025-12-26 |
2,072 |
5 |
1 |
1 |
1 |
3 |
| On the first day of #Shipmas, Roboflow gave to me SAM3 in 3d.
Meta's new model that can turn any object in a photo int… |
@roboflow |
Company |
Original |
2025-12-25 |
6,777 |
84 |
7 |
2 |
1 |
30 |
| ’Twas the night before #Shipmas
When all through the cloud,
Not a model was training, not a pipeline was loud.
The GPU… |
@roboflow |
Company |
Original |
2025-12-24 |
6,402 |
24 |
0 |
2 |
1 |
7 |
| deploying to the edge just got way easier
run vision models in minutes on a jetson orin nano
hardware like this is wh… |
@roboflow |
Company |
Quote |
2025-12-19 |
3,698 |
47 |
6 |
3 |
0 |
25 |
| it's been 5 days since @measure_plan made like 10 viral videos showing that building with visual artificial intelligenc… |
@roboflow |
Company |
Quote |
2025-12-15 |
22,242 |
356 |
26 |
8 |
1 |
331 |
| this app is cool on mondays because folks that build stuff on the weekends share their work
automating notifications … |
@roboflow |
Company |
Quote |
2025-12-15 |
6,528 |
109 |
4 |
3 |
1 |
37 |
| OAK 4 device is unlocking visual artificial intelligence at the edge with better hardware
might be the easiest way to … |
@roboflow |
Company |
Quote |
2025-12-12 |
3,799 |
35 |
8 |
0 |
0 |
30 |
| vibe code the app
vibe train the model
get fit and/or rich in 1 hr
welcome to the future |
@roboflow |
Company |
Quote |
2025-12-12 |
5,110 |
50 |
3 |
2 |
0 |
37 |
| if you have only been in LLM land for the past few years, probably time to give pixels a shot
we are bringing the barr… |
@roboflow |
Company |
Quote |
2025-12-11 |
4,140 |
49 |
1 |
3 |
0 |
19 |
| oh come on when is the last time an open source tutorial gave you an output like this
a round of applause for @skalski… |
@roboflow |
Company |
Quote |
2025-12-10 |
19,044 |
336 |
20 |
3 |
0 |
208 |
| Announcing Roboflow Rapid: the first prompt-based model creation engine. Data labeling is dead.
Go from idea to deploy… |
@roboflow |
Company |
Original |
2025-12-09 |
22,992 |
151 |
13 |
12 |
7 |
121 |
| Vision arena and model playground helping people realize that visual artificial intelligence made a huge leap in 2025
… |
@roboflow |
Company |
Quote |
2025-12-05 |
3,727 |
44 |
6 |
1 |
0 |
30 |
| The most time intensive open source tutorial on our Youtube channel
Step by step, all code open source, datasets avail… |
@roboflow |
Company |
Quote |
2025-12-04 |
30,146 |
584 |
55 |
11 |
0 |
412 |
| open source foundation models will set you free |
@roboflow |
Company |
Quote |
2025-12-04 |
10,457 |
187 |
7 |
2 |
0 |
69 |
| Gemini 3 added to Roboflow Playground, results starting to come in
SAM 3 looking like it's still the zero-shot detect… |
@roboflow |
Company |
Quote |
2025-12-02 |
3,133 |
25 |
2 |
0 |
0 |
6 |
| Gemini 3 is topping lots of charts
It just topped the vision vibe benchmark for LLMs
Time to at least start testing G… |
@roboflow |
Company |
Quote |
2025-12-01 |
1,070 |
2 |
0 |
1 |
0 |
2 |
| people build cool stuff when new models improve the state of the art
👏👏👏👏👏 @measure_plan |
@roboflow |
Company |
Quote |
2025-12-01 |
1,945 |
13 |
0 |
1 |
0 |
5 |
| a good example of how broad SAM 3 can go is RF100-VL and ODinW13 benchmarks
both use real world datasets across domai… |
@roboflow |
Company |
Quote |
2025-11-21 |
3,574 |
28 |
2 |
3 |
0 |
14 |
| ok the blueprint is now clear on how to train a breakthrough model and get viral adoption
have you been paying attenti… |
@roboflow |
Company |
Quote |
2025-11-20 |
3,229 |
24 |
1 |
2 |
0 |
5 |
| we are going live tomorrow to get hands on with SAM 3
lots of demos and ongoing Q&A, join in |
@roboflow |
Company |
Quote |
2025-11-20 |
8,726 |
90 |
8 |
1 |
0 |
34 |
| SAM 3 is the missing link for vision to be as easy to use as LLMs
completely changes how people will build computer vi… |
@roboflow |
Company |
Quote |
2025-11-19 |
87,621 |
894 |
68 |
9 |
4 |
516 |
| In other important AI breakthrough news: https://t.co/wHZnqTsU3o |
@roboflow |
Company |
Original |
2025-11-18 |
1,282 |
5 |
1 |
1 |
0 |
0 |
| Gemini 3 is insanely good at a lot of things
not yet beating a 7 yr old on some tasks tho
keep that in mind https://t… |
@roboflow |
Company |
Quote |
2025-11-18 |
2,491 |
18 |
1 |
9 |
0 |
0 |
| latest supervision release is the missing link to help people see the power of Qwen3-VL
if you're not convinced open s… |
@roboflow |
Company |
Quote |
2025-11-18 |
1,381 |
13 |
0 |
0 |
0 |
3 |
| Segment Anything URL has a countdown clock that ends on 19th
May I meet you? https://t.co/rxnafs0Yn6 |
@roboflow |
Company |
Original |
2025-11-17 |
5,617 |
72 |
6 |
4 |
0 |
25 |
| rf-detr-seg has the speed and accuracy combination that's been missing for sports applications |
@roboflow |
Company |
Quote |
2025-11-17 |
16,115 |
236 |
19 |
2 |
1 |
115 |
| this took me only 1 minute to create.
i did not apply myself and i did not work hard, but this is insanely easy https:… |
@roboflow |
Company |
Quote |
2025-11-16 |
4,261 |
28 |
3 |
3 |
1 |
18 |
| RF-DETR is now state of the art for object detection and segmentation
The segmentation graph makes the YOLO vs transfo… |
@roboflow |
Company |
Quote |
2025-11-16 |
32,919 |
363 |
36 |
6 |
4 |
170 |
| GPT-5.1 is hitting the same 75.9% ceiling on the Vision AI Checkup evaluation as O4 Mini, GPT-5, and GPT-5 Mini https:/… |
@roboflow |
Company |
Quote |
2025-11-14 |
2,029 |
11 |
1 |
0 |
0 |
1 |
| benchmarks for VLMs across many different datasets to help you find the best one for your use case
https://t.co/SmGvi0… |
@roboflow |
Company |
Quote |
2025-11-07 |
1,234 |
4 |
1 |
0 |
0 |
3 |
| sports are a great way to get into vision AI. lots of hard but fun problems to work out.
checkout our sports repo to g… |
@roboflow |
Company |
Quote |
2025-11-05 |
1,112 |
10 |
0 |
1 |
0 |
5 |
| new youtube video incoming |
@roboflow |
Company |
Quote |
2025-11-05 |
907 |
4 |
0 |
0 |
0 |
1 |
| idk what other type of content someone would want to see on this website? |
@roboflow |
Company |
Quote |
2025-10-29 |
1,217 |
4 |
0 |
1 |
0 |
2 |
| At @github Universe's "mini MCP" summit, we break down how MCP and 200k+ open source vision models enable agents to que… |
@roboflow |
Company |
Original |
2025-10-29 |
1,331 |
13 |
3 |
3 |
1 |
5 |
| we 💜 RMFG
(And added more free GPU to your account. Let it rip) |
@roboflow |
Company |
Quote |
2025-08-23 |
1,978 |
12 |
0 |
1 |
0 |
3 |
| RF-DETR, SOTA for object detection, was built by a team of 3
join the livestream to hear how we pulled that off |
@roboflow |
Company |
Quote |
2025-08-19 |
5,810 |
75 |
7 |
1 |
0 |
15 |
| great initiative to get RF-DETR into RKNN Model Zoo, join in if this is something you'd want to build with |
@roboflow |
Company |
Quote |
2025-08-19 |
1,173 |
5 |
1 |
0 |
0 |
1 |
| amazing how results can change by just using two open source libraries with the same model |
@roboflow |
Company |
Quote |
2025-08-18 |
4,227 |
65 |
9 |
0 |
0 |
26 |
| getting close to releasing the full video
everything open source and available for you to build with |
@roboflow |
Company |
Quote |
2025-08-13 |
8,861 |
67 |
7 |
1 |
0 |
29 |
| visual AI running on glasses is the future
zuck called this out in his personal super intelligence memo yesterday http… |
@roboflow |
Company |
Quote |
2025-07-31 |
1,227 |
9 |
3 |
0 |
0 |
3 |
| autonomous metal fabrication factory being built
robotics + AI unlocks huge opportunities |
@roboflow |
Company |
Quote |
2025-07-31 |
762 |
4 |
0 |
0 |
0 |
2 |
| we basically did the same thing with DINOv2 and O365 to make RF-DETR |
@roboflow |
Company |
Quote |
2025-07-30 |
1,275 |
12 |
1 |
1 |
0 |
1 |
| new object detection leaderboard is live
we didn't mean to time this release with paperswithcode being sunset today
l… |
@roboflow |
Company |
Quote |
2025-07-25 |
1,394 |
12 |
0 |
1 |
0 |
6 |
| real-time AI use cases for wearables and mobile are now possible
we released RF-DETR nano, small and medium to give d… |
@roboflow |
Company |
Quote |
2025-07-24 |
56,007 |
496 |
62 |
8 |
8 |
323 |
| Audi and Ogilvy used Roboflow to turn your phone into a high context car manual
point at a feature, learn about it |
@roboflow |
Company |
Quote |
2025-07-23 |
820 |
6 |
0 |
0 |
0 |
0 |
| everything you see in this video snippet, we give away for free in the blog post
you can just do things |
@roboflow |
Company |
Quote |
2025-07-22 |
2,012 |
13 |
3 |
1 |
0 |
2 |
| supervision is #3 trending repo on GitHub today and almost 30k stars
give it a try and see what the hype is about http… |
@roboflow |
Company |
Quote |
2025-07-22 |
1,715 |
7 |
1 |
1 |
0 |
2 |
| The first Mentra AR glasses hackathon took place last weekend at @ycombinator
Mentra dropped a first version of their… |
@roboflow |
Company |
Original |
2025-07-17 |
2,310 |
34 |
3 |
1 |
1 |
9 |
| RF-DETR release next week
more SOTA models coming in smaller sizes, all open source and Apache 2.0 |
@roboflow |
Company |
Quote |
2025-07-17 |
83,887 |
522 |
95 |
9 |
3 |
293 |
| 🦝🦝🦝 |
@roboflow |
Company |
Quote |
2025-07-09 |
805 |
2 |
0 |
0 |
0 |
0 |
| 💜💜💜
check your account for more free credits we added. happy deploying 🫡 |
@roboflow |
Company |
Quote |
2025-06-26 |
1,576 |
6 |
1 |
2 |
1 |
0 |
| You can now train models with @huggingface's SmolVLM2 and SmolVLM 256M architectures on Roboflow 📈
Learn how to train … |
@roboflow |
Company |
Original |
2025-06-23 |
4,223 |
37 |
5 |
0 |
0 |
15 |
| $150k in Roboflow credits for the 40 teams building with AI.
Apply before May 17! |
@roboflow |
Company |
Quote |
2025-05-16 |
2,688 |
19 |
3 |
1 |
0 |
5 |
| New few-shot object detection challenge for CVPR this year, can foundation models replace manual annotations for 2D ima… |
@roboflow |
Company |
Quote |
2025-04-23 |
1,439 |
10 |
0 |
1 |
0 |
4 |
| OpenAI o3 and o4-mini multimodal and vision analysis is now available as well. Both models failed at simple object coun… |
@roboflow |
Company |
Quote |
2025-04-17 |
2,208 |
15 |
2 |
1 |
0 |
1 |
| We ran GPT-4.1 on a series of vision tests to see how it performs.
Base model and nano both do just as well as GPT-4.… |
@roboflow |
Company |
Original |
2025-04-15 |
4,897 |
23 |
4 |
3 |
1 |
6 |
| First release after launching RF-DETR is out, let us know what should be in the upcoming release or join in and contrib… |
@roboflow |
Company |
Quote |
2025-04-03 |
1,137 |
14 |
0 |
0 |
0 |
2 |
| Train and deploy RF-DETR with Roboflow is now live.
This is just the start. More coming next week.
The craziest RF-… |
@roboflow |
Company |
Quote |
2025-03-27 |
1,013 |
4 |
0 |
0 |
0 |
2 |
| Join the livestream happening now for live coding and Q&A around the new RF-DETR model with @skalskip92
https://t… |
@roboflow |
Company |
Quote |
2025-03-27 |
1,603 |
9 |
0 |
0 |
0 |
1 |
| https://t.co/EvBEh3x9wV |
@roboflow |
Company |
Original |
2025-03-26 |
727 |
7 |
0 |
1 |
0 |
0 |
| More about why we built RF-DETR from our CEO @josephofiowa
And the benchmark we developed alongside the model, RF100-… |
@roboflow |
Company |
Quote |
2025-03-20 |
1,340 |
9 |
1 |
1 |
0 |
0 |
| Excited to announce RF-DETR, the current SOTA for real-time object detection, fully open source and Apache 2.0 for the … |
@roboflow |
Company |
Quote |
2025-03-20 |
191,935 |
1,609 |
259 |
19 |
16 |
1,195 |
| Exciting new edge devices for the community
OpenMV N6 & AE3 - Low Power, Python Programmable AI Cameras
All purc… |
@roboflow |
Company |
Quote |
2025-03-18 |
1,451 |
12 |
1 |
1 |
0 |
2 |
| AGI has been delivered just in time for the weekend |
@roboflow |
Company |
Quote |
2025-03-14 |
7,039 |
20 |
2 |
1 |
1 |
11 |
| Livestream starting in 15 minutes!
Learn about the latest zero-shot model and how to fine-tune it for your next projec… |
@roboflow |
Company |
Quote |
2025-03-14 |
2,374 |
5 |
0 |
0 |
0 |
1 |
| Building a robot from scratch in a weekend is now possible.
Robotics + AI + builders are going to change everything ab… |
@roboflow |
Company |
Quote |
2025-02-12 |
3,819 |
37 |
5 |
1 |
0 |
5 |
| Great to see the object detection leaderboard being helpful to understand performance across models, sizes, and license… |
@roboflow |
Company |
Quote |
2025-02-11 |
3,190 |
15 |
4 |
1 |
0 |
8 |
| Need potholes datasets?
search Universe: https://t.co/V2NQAPpyX2 |
@roboflow |
Company |
Quote |
2025-02-06 |
2,346 |
15 |
2 |
1 |
0 |
3 |
| please help them |
@roboflow |
Company |
Quote |
2025-02-03 |
1,262 |
12 |
0 |
0 |
0 |
0 |
| You can build AI systems like this using a fully open-source library, completely free. And it’s easier than you think.
… |
@roboflow |
Company |
Original |
2025-01-30 |
2,333 |
43 |
6 |
2 |
0 |
27 |
| The key to getting high-quality generative AI outputs is using computer vision to guide and refine what your models gen… |
@roboflow |
Company |
Original |
2025-01-30 |
563 |
5 |
0 |
2 |
0 |
3 |
| It's hard to know if GPT-4o, Claude, or Gemini is best for what you're working on.
We built a free workflow to compar… |
@roboflow |
Company |
Original |
2025-01-29 |
587 |
4 |
0 |
1 |
0 |
1 |
| Ever wondered how player movements can turn into 3D visualization? It starts with pose estimation to detect keypoints a… |
@roboflow |
Company |
Original |
2025-01-29 |
755 |
8 |
1 |
1 |
0 |
11 |
| Fine-tune your own background removal app in a few minutes. Isolate objects in images, whether for creative projects li… |
@roboflow |
Company |
Original |
2025-01-27 |
440 |
2 |
0 |
1 |
0 |
1 |
| 🧵 How do self-driving cars see?
Semantic segmentation can be used to label every pixel in an image. This helps differen… |
@roboflow |
Company |
Original |
2025-01-22 |
1,273 |
4 |
0 |
1 |
0 |
5 |
| Imagine an AI camera operator that keeps the focus exactly where you want it—tracking players, wildlife, security, or K… |
@roboflow |
Company |
Original |
2025-01-21 |
633 |
3 |
0 |
1 |
0 |
4 |
| 🧵 Spotting defects early in manufacturing can save you from expensive downstream issues. Automated defect detection sys… |
@roboflow |
Company |
Original |
2025-01-18 |
639 |
1 |
0 |
1 |
0 |
1 |
| 🧵 Small object detection in high-res images is tricky. InferenceSlicer solves this by slicing images, detecting in each… |
@roboflow |
Company |
Original |
2025-01-17 |
483 |
9 |
0 |
1 |
0 |
6 |
| 🧵 The fastest way to label image data for fine-tuning is by automating it with a foundation model. Auto Label, built on… |
@roboflow |
Company |
Original |
2025-01-16 |
753 |
9 |
1 |
1 |
0 |
5 |
| 🧵 Most foundation models don’t handle private data or niche cases well, but fine-tuning is easier than you think. PaliG… |
@roboflow |
Company |
Original |
2025-01-15 |
608 |
4 |
0 |
2 |
1 |
0 |
| Building models is step one—what comes next is even more exciting.
Here's how we think about creating business impact … |
@roboflow |
Company |
Original |
2025-01-14 |
413 |
4 |
0 |
1 |
0 |
0 |
| Florence-2 is an open-sourced vision-language model (VLM) by Microsoft, useful in tasks like captioning, object detecti… |
@roboflow |
Company |
Original |
2025-01-13 |
381 |
2 |
0 |
1 |
0 |
1 |
| AI models to real-world actions, simplified.
Try it now with Roboflow Workflows. https://t.co/z8EmWeQCaq |
@roboflow |
Company |
Original |
2025-01-11 |
547 |
1 |
0 |
0 |
0 |
0 |
| People tend to use computer vision applications for monitoring but monitoring isn't useful if you aren't alerted to imp… |
@roboflow |
Company |
Original |
2025-01-10 |
523 |
4 |
0 |
2 |
0 |
0 |
| SAM-2.1 (Segment Anything Model 2.1) is a segmentation model from Meta Research that improves on its predecessor, SAM-2… |
@roboflow |
Company |
Original |
2025-01-09 |
312 |
3 |
0 |
1 |
0 |
3 |
| Fast solutions for multimodal use cases, from idea to production—this changes everything. https://t.co/Tts09wPvy7 |
@roboflow |
Company |
Original |
2025-01-08 |
531 |
3 |
1 |
0 |
0 |
5 |
| RT-DETR (Real-Time DEtection TRansformer) is a transformer-based object detection model developed by Peking University … |
@roboflow |
Company |
Original |
2025-01-08 |
641 |
10 |
0 |
1 |
0 |
4 |
| How do Embeddings work?
Embeddings convert images and text into numerical representations called vectors, allowing comp… |
@roboflow |
Company |
Original |
2025-01-07 |
542 |
13 |
0 |
1 |
0 |
7 |
| You’ve trained a vision AI model, now what? Join tomorrow's live session with a Roboflow engineer and learn how to put … |
@roboflow |
Company |
Original |
2025-01-07 |
508 |
9 |
0 |
0 |
0 |
2 |
| Getting vision models to structure tables of information just got a lot easier. You can build a workflow that uses Open… |
@roboflow |
Company |
Original |
2025-01-06 |
734 |
8 |
0 |
1 |
1 |
4 |
| If you’re building a computer vision model, the quality of your dataset defines the quality of your results.
The secre… |
@roboflow |
Company |
Original |
2025-01-03 |
1,301 |
13 |
3 |
1 |
0 |
11 |
| What used to take a team experts to build can now be done in minutes using existing models.
Tons of opportunity ahead … |
@roboflow |
Company |
Original |
2025-01-03 |
459 |
2 |
0 |
0 |
0 |
1 |
| Choosing between Vision Transformers (ViTs) and Convolutional Neural Networks (CNNs) for object detection depends on yo… |
@roboflow |
Company |
Original |
2025-01-02 |
583 |
7 |
0 |
1 |
0 |
2 |
