Cognition
@cognition · San Francisco Bay Area · joined 29 Jan 2024
Makers of Devin, the first AI software engineer. We are an applied AI lab building end-to-end software agents. Join us: https://t.co/4Ss9hvpjRG
184 987Followers
8Following
973Posts total
12.4MViews on collected posts
Latest posts
Looking for new posts…
@cognition Need it in Fusion :( https://t.co/cL9cZ0f9Rg
638 views · 7 likes
· 0 reposts · 0 replies
Open on X →
Try it in Devin Desktop and Devin CLI:
https://t.co/y3humFMnN0
3.1K views · 11 likes
· 0 reposts · 1 replies
Open on X →
On FrontierCode 1.1, our benchmark for real-world engineering tasks that grades mergeability and quality, Opus 5.5 scores 65.3% on Extended — taking the leading spot from Fable 5 at a fraction of the cost.
Read more: https://t.co/ELoIAbeica
4.7K views · 19 likes
· 1 reposts · 1 replies
Open on X →
Claude Opus 5.5 is now available in Devin.
On FrontierCode 1.1, Opus 5.5 takes the #1 sport from Fable 5 at a fraction of the cost. https://t.co/QUhqpGjO9X
75.2K views · 579 likes
· 34 reposts · 44 replies
Open on X →
@cognition Impressive! How did the team clear Apple Legal to serve more than two macOS VMs per host?
https://t.co/1JH5y6MHCY https://t.co/BpxjuMy5xW
5.4K views · 28 likes
· 1 reposts · 4 replies
Open on X →
Read more about how we built this:
https://t.co/gaRMfdv1gA
33.5K views · 140 likes
· 7 reposts · 3 replies
Open on X →
Try it out in Devin:
https://t.co/Z1eWTVRBwn
20.3K views · 62 likes
· 2 reposts · 1 replies
Open on X →
Special delivery: Devin just got a Mac 🍎
Now Devin can:
1. Build & test apps on its own Mac VM with iOS simulator
2. Send a screen recording via Slack
3. Send a TestFlight link so you can start using it 📲 https://t.co/zpOGD3eTV9
2M views · 3.3K likes
· 303 reposts · 188 replies
Open on X →
0:45
@cognition 0.5% worse than fable for 36% cheaper. insane
2.1K views · 32 likes
· 0 reposts · 0 replies
Open on X →
Try it out in Devin CLI:
https://t.co/ErdEXYPKhz https://t.co/20EMbzPfbN
8.5K views · 57 likes
· 1 reposts · 2 replies
Open on X →
Read more in our blog:
https://t.co/AhYs1S1mrR
7.4K views · 48 likes
· 1 reposts · 0 replies
Open on X →
Unlike plain model routing, Fusion keeps frontier intelligence in charge.
We don’t rely on one initial routing decision. The lead reviews work, flags problems, and can take control back when the sidekick is out of its depth. https://t.co/4d9mRVkf2M
19.1K views · 99 likes
· 0 reposts · 6 replies
Open on X →
We partnered with @ArtificialAnlys and @ValsAI to evaluate Fusion across several coding agent benchmarks.
The cost savings hold across the board while maintaining frontier performance. https://t.co/CeffWEUwQ0
39K views · 119 likes
· 5 reposts · 6 replies
Open on X →
Introducing Fusion in Devin CLI
The most efficient frontier harness for Fable & Astra; 39% cheaper across coding benchmarks.
Pick your favorite model for planning and a cost-effective model for execution. https://t.co/rKVn6S7Hqu
351.6K views · 1.5K likes
· 120 reposts · 112 replies
Open on X →
@cognition wow! what about terminalbench 4.0 score
15.5K views · 181 likes
· 0 reposts · 4 replies
Open on X →
SWE-2 is available today in Devin across Desktop and CLI. We’re making it free for all Pro, Max & Teams subscribers for the next month.
Read more about how we trained SWE-2:
https://t.co/s5L08Rmxgi
125.4K views · 508 likes
· 28 reposts · 36 replies
Open on X →
As a result, SWE-2 is our most capable model yet for software engineering tasks – while achieving market-leading price performance.
Our team at Cognition uses it for feature development, debugging, and even complex visualizations of novel math:
https://t.co/N5kLeaiQ7Z
64.3K views · 252 likes
· 2 reposts · 1 replies
Open on X →
The result is a model that is way more efficient than SWE-1.7.
On FrontierCode 1.1 Main, SWE-2 medium scores higher than SWE-1.7 while taking 58% fewer turns and costing 81% less on average.
We observe that SWE-2 learns to explore in a more focused way before making changes. ht
46.9K views · 307 likes
· 5 reposts · 3 replies
Open on X →
SWE-2 is our first model to support effort levels. For that, we made significant improvements to our length penalty recipe - read more in our blog post!
In a single RL run, we are able to push the Pareto curve while preserving its shape: medium effort becomes both cheaper & htt
80K views · 445 likes
· 11 reposts · 1 replies
Open on X →
GIF
SWE-2 is post-trained on Kimi-K3 and proves that our RL recipe continues to scale on stronger base models.
On FrontierCode, SWE-2 achieves a score of 50.0%. It beats SWE-1.7, Grok 4.6, and GPT 5.6 Sol while matching Fable 5.1 at 64% lower cost. https://t.co/O8dARqVRhM
116.6K views · 824 likes
· 33 reposts · 12 replies
Open on X →
Introducing SWE-2, our closest model yet to the frontier.
On leading evals, it scores on par with recent frontier models – at up to 70% lower cost.
We scaled RL to multiple trillions of parameters, with a refined recipe that pushes the Pareto curve on both capabilities & cost.
2.1M views · 6.6K likes
· 492 reposts · 427 replies
Open on X →
a visualization of the flow field of the navier-stokes finite-time blowup solution. it's a good model sir
(singularity at t = 1 with viscosity matching water, unit distance ~ 1 mm. leading and first order background terms)
https://t.co/PkjGXaizNC https://t.co/lrQDYjUoBG
168.7K views · 586 likes
· 36 reposts · 17 replies
Open on X →
0:18
In the next chapter of software engineering, agents will become proactive by default, software will improve itself, and compute will automatically get allocated toward the highest impact use cases.
We’re still at the dawn of the self-driving software era.
Read more:
14.4K views · 113 likes
· 4 reposts · 8 replies
Open on X →
Delivering productive engineering value takes a deep understanding of each customer's systems and processes.
To work side by side with teams worldwide, we've opened offices in D.C., Tokyo, Singapore, London, São Paulo, and Madrid in the past year.
14K views · 97 likes
· 0 reposts · 2 replies
Open on X →
Devin is taking on more directed, proactive work.
With automations, Devin watches Slack channels for alerts, then autonomously investigates and remediates issues.
Devin Security Swarm scans, maps, and reduces large codebases to find and triage multi-step exploits.
15.1K views · 107 likes
· 0 reposts · 2 replies
Open on X →
As Devin becomes essential to how companies build software, customers across the economy are rapidly expanding their use – from chip design at @nvidia and aviation at @GE_Aerospace to financial services at @Citi, automotive at @MercedesBenz, and AI infrastructure at @modal. https
22.5K views · 162 likes
· 4 reposts · 3 replies
Open on X →
Other investors in the round include Benchmark, Bessemer, Kleiner Perkins, Greylock, Lightspeed, Altimeter, Bond Capital, Meritech, Atreides, Valor, T. Rowe Price, Lux, 8VC, D1, 137 Ventures, DST, Positive Sum, HOF, Soma, BoxGroup, Ribbit Capital, Swish, Stripes, Definition,
51.7K views · 189 likes
· 3 reposts · 5 replies
Open on X →
The world needs far more software than it can build. Cognition exists to change that.
We’ve just raised over $2B at a $48B valuation, led by a16z, Accel, Founders Fund, General Catalyst, and Avenir.
Since our round in May, run-rate revenue has grown from $492 M to almost $900 h
1.4M views · 2.4K likes
· 246 reposts · 210 replies
Open on X →
@cognition 64% cheaper and nobody is gonna miss the 0.4 points
2.4K views · 6 likes
· 0 reposts · 0 replies
Open on X →
Against accounts of the same size
46 posts from the last 90 days, next to the 100K–1M follower range. shown widely, but few of those viewers react.
Median views15 308this account5 983median for 100K–1M
Reach, %8.27%this account1.49%median for 100K–1M
Engagement, %0.49%this account0.93%median for 100K–1M
| Metric | This account | Median for 100K–1M | Ratio |
|---|---|---|---|
| Median views per post | 15 308 | 5 983 | 2.56× |
| Reach (views ÷ followers) | 8.27% | 1.49% | 5.55× |
| Engagement rate | 0.49% | 0.93% | 0.52× |
Others in this range → Compare with another account → How these benchmarks are built →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
351.6K11 Sep
39K
19.1K
7.4K
8.5K
2.1K
2M15 Sep
20.3K
33.5K
5.4K
75.2K22 Sep
4.7K
3.1K
638
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
0.52%11 Sep
0.34%
0.57%
0.66%
0.72%
1.50%
0.20%15 Sep
0.32%
0.46%
0.61%
0.91%22 Sep
0.45%
0.39%
1.10%
Reactions — likes, reposts, replies and quotes — divided by views. Median for 100K–1M accounts is 0.93%.
What the audience does
Likes69.6%18 878 in total
Reposts4.9%1 342 in total
Replies4.1%1 118 in total
Quotes2.3%634 in total
Bookmarks19.0%5 167 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Similar accounts
HDFC Bank Cares
@HDFCBank_Cares
370K followers
· 1.7M posts
Lekau Sehoana ✓
@LekauSehoana
369.9K followers
· 22K posts
Anders Holm
@ders808
369.9K followers
· 21.3K posts
fluxpavilion.wav
@Fluxpavilion
369.9K followers
· 10.7K posts
Texas A&M Football ✓
@AggieFootball
369.8K followers
· 29.3K posts
Jargon ✓
@JargonMedya
369.8K followers
· 39.2K posts
Academy of American Poets
@POETSorg
369.8K followers
· 46K posts
T1 VALORANT ✓
@T1
369.8K followers
· 10.3K posts