Cognition(@cognition_labs)

Introducing Fusion in Devin CLI The most efficient frontier harness for Fable & Astra; 39% chea...

Cognition (@cognition) — Fri, 11 Sep 2026 16:16:23 GMT

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.

💬60🔄62❤️869👀71232📊181

Read the docs: https://t.co/vNxplaLUzC

Cognition (@cognition) — Thu, 10 Sep 2026 20:35:37 GMT

Read the docs:

docs.devin.ai/work-with-devi…

💬3🔄1❤️31👀7214📊7

Try it today in Devin: https://t.co/wfs9rGAyt2

Cognition (@cognition) — Thu, 10 Sep 2026 20:12:53 GMT

Try it today in Devin:

app.devin.ai

💬7🔄2❤️41👀10083📊11

Introducing Devin Voice 🦦 ☎️ Your favorite AI software engineer just got a landline. You say it...

Cognition (@cognition) — Thu, 10 Sep 2026 20:12:52 GMT

Introducing Devin Voice 🦦 ☎️ 

Your favorite AI software engineer just got a landline. 

You say it, Devin ships it.

Powered by GPT-Live and our new SWE-2 model.

Your browser does not support the video tag.

Cognition@cognition

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.

💬109🔄123❤️1472👀278105📊272

Welcome @jkelleyrtp and the entire Dioxus team to Cognition! Dioxus is one of the most beloved open...

Cognition (@cognition) — Thu, 10 Sep 2026 17:59:30 GMT

Welcome @jkelleyrtp and the entire Dioxus team to Cognition!

Dioxus is one of the most beloved open source Rust frameworks. We're proud to continue supporting Dioxus, Blitz, Taffy, and Subsecond while bringing the team's expertise to Devin's VM, computer use, and testing.

Jon Kelley@jkelleyrtp

We have some very big news to share. 🎉

Today, Dioxus Labs is joining @Cognition to accelerate the development of Devin, Cognition’s autonomous cloud coding agent.

We are excited to continue building Dioxus while also helping Cognition shape the future of software engineering.

Your browser does not support the video tag.

💬20🔄24❤️278👀37190📊41

We've used Dioxus extensively to build Devin. When the two teams met, the fit was obvious. Jonathan...

Cognition (@cognition) — Thu, 10 Sep 2026 17:59:30 GMT

We've used Dioxus extensively to build Devin. When the two teams met, the fit was obvious.

Jonathan and the Dioxus team are exceptional engineers and open source builders. At Cognition, we love working with founders.

More on why we're so excited: cognition.com/blog/welcoming…

💬1🔄0❤️16👀3763📊3

As a result, SWE-2 is our most capable model yet for software engineering tasks – while achieving ma...

Cognition (@cognition) — Thu, 10 Sep 2026 15:21:03 GMT

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:

x.com/penlume/status…

penlu@penlume

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)

x.com/OpenAI/status/…

Your browser does not support the video tag.

💬1🔄2❤️241👀56687📊19

SWE-2 is available today in Devin across Desktop and CLI. We’re making it free for all Pro, Max &...

Cognition (@cognition) — Thu, 10 Sep 2026 15:21:03 GMT

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

💬36🔄26❤️482👀117597📊108

SWE-2 is post-trained on Kimi-K3 and proves that our RL recipe continues to scale on stronger base m...

Cognition (@cognition) — Thu, 10 Sep 2026 15:21:01 GMT

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.

💬9🔄30❤️774👀96335📊99

SWE-2 is our first model to support effort levels. For that, we made significant improvements to our...

Cognition (@cognition) — Thu, 10 Sep 2026 15:21:01 GMT

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 & smarter; max effort learns to use more tokens & turns to achieve the highest scores.

💬1🔄8❤️352👀41856📊37

Introducing SWE-2, our closest model yet to the frontier. On leading evals, it scores on par with r...

Cognition (@cognition) — Thu, 10 Sep 2026 15:21:00 GMT

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.

💬395🔄444❤️6110👀1706979📊1279

You might have seen it published here: https://t.co/PuEWqn83rv

Cognition (@cognition) — Wed, 09 Sep 2026 19:55:48 GMT

You might have seen it published here:

x.com/penlume/status…

penlu@penlume

4397328654844826923795068102505872571721883526553349659561256924505973939597593482272505698004801207988043088656411102133523080581 divides RSA-260

💬1🔄0❤️58👀18008📊6

If this type of research interests you, we encourage you to consider joining Cognition: https://t.c...

Cognition (@cognition) — Wed, 09 Sep 2026 19:55:48 GMT

If this type of research interests you, we encourage you to consider joining Cognition:

cognition.ai/careers

💬1🔄0❤️46👀6034📊7

Last week we published a factorization of RSA-260. Today, we’re sharing the methodology of how Dev...

Cognition (@cognition) — Wed, 09 Sep 2026 19:55:47 GMT

Last week we published a factorization of RSA-260. 

Today, we’re sharing the methodology of how Devin and a Cognition researcher built the world’s fastest GPU optimized lattice siever, to make factoring numbers 10x cheaper than the previous state of the art:

cognition.com/blog/factoring…

💬25🔄98❤️776👀198160📊168

Delivering productive engineering value takes a deep understanding of each customer's systems and pr...

Cognition (@cognition) — Tue, 08 Sep 2026 17:01:44 GMT

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.

💬2🔄0❤️97👀13693📊10

In the next chapter of software engineering, agents will become proactive by default, software will ...

Cognition (@cognition) — Tue, 08 Sep 2026 17:01:44 GMT

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: cognition.com/blog/series-e

💬8🔄4❤️112👀14055📊20

Devin is taking on more directed, proactive work. With automations, Devin watches Slack channels f...

Cognition (@cognition) — Tue, 08 Sep 2026 17:01:43 GMT

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.

💬2🔄0❤️107👀14467📊13

Other investors in the round include Benchmark, Bessemer, Kleiner Perkins, Greylock, Lightspeed, Alt...

Cognition (@cognition) — Tue, 08 Sep 2026 17:01:42 GMT

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, Hanabi, Journey, Bain Capital Ventures, Layer Global, A* Capital, Alpha Wave, Alkeon, Diffusion, and NVIDIA.

💬4🔄3❤️188👀48901📊27

As Devin becomes essential to how companies build software, customers across the economy are rapidly...

Cognition (@cognition) — Tue, 08 Sep 2026 17:01:42 GMT

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.

💬3🔄4❤️160👀22103📊18

The world needs far more software than it can build. Cognition exists to change that. We’ve just ra...

Cognition (@cognition) — Tue, 08 Sep 2026 17:01:41 GMT

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 M.

💬205🔄238❤️2407👀1340587📊558

Access to Astra is rolling out over the coming days. Starting today, enterprise customers in OpenAI...

Cognition (@cognition) — Thu, 03 Sep 2026 19:40:15 GMT

Access to Astra is rolling out over the coming days.

Starting today, enterprise customers in OpenAI's Daybreak Program can get access to Astra.

Read more on our blog: 

devin.ai/blog/gpt-6-ast…

💬0🔄3❤️74👀18616📊8

GPT-6 Astra is coming to Devin. On FrontierCode 1.1, Astra performs within 0.4 points of Fable 5 at...

Cognition (@cognition) — Thu, 03 Sep 2026 19:40:14 GMT

GPT-6 Astra is coming to Devin.

On FrontierCode 1.1, Astra performs within 0.4 points of Fable 5 at a 64% lower cost. It also sets a new SOTA on our internal testing benchmark, generating more comprehensive tests, clearer reports, and better video evidence.

💬88🔄107❤️1293👀845712📊252

Try it out in Devin: https://t.co/wfs9rGAyt2

Cognition (@cognition) — Tue, 01 Sep 2026 18:06:51 GMT

Try it out in Devin:

app.devin.ai

💬0🔄0❤️14👀4140📊2

In Devin Fusion, we pair frontier models for planning with cost-effective models for execution. Wit...

Cognition (@cognition) — Tue, 01 Sep 2026 18:06:50 GMT

In Devin Fusion, we pair frontier models for planning with cost-effective models for execution.

With Fable 5.1, this orchestration performs even better. We highly recommend Fusion as a cost-effective way to access frontier intelligence.

💬2🔄4❤️70👀26557📊12

Introducing Fable 5.1 in Devin Fable-level intelligence is now 54% cheaper – making it even cheaper...

Cognition (@cognition) — Tue, 01 Sep 2026 18:06:49 GMT

Introducing Fable 5.1 in Devin

Fable-level intelligence is now 54% cheaper – making it even cheaper than Opus – due to a change in caching.

Devin’s Fusion harness is now even smarter and cheaper than before, matching Fable 5.1 on FrontierCode at 47% lower cost. Here's how:

💬68🔄77❤️1123👀214576📊212

Over 95% of tokens from a coding task are cached tokens. Fable 5.1 dropped the price of cached toke...

Cognition (@cognition) — Tue, 01 Sep 2026 18:06:49 GMT

Over 95% of tokens from a coding task are cached tokens.

Fable 5.1 dropped the price of cached tokens by 4x, making it much more efficient for coding as a whole, despite higher uncached prices.

💬7🔄7❤️111👀23008📊22

Get your slippers on and join us for a walk around our SF headquarters, where we're doing it all wit...

Cognition (@cognition) — Fri, 28 Aug 2026 15:34:14 GMT

Get your slippers on and join us for a walk around our SF headquarters, where we're doing it all with Devin. Even ordering our morning coffee.

cognition.com/careers

Your browser does not support the video tag.

💬48🔄60❤️988👀589480📊210

Try it today at https://t.co/0Su90EOTkG

Cognition (@cognition) — Wed, 26 Aug 2026 16:00:23 GMT

Try it today at devin.ai

💬0🔄0❤️2👀560

We also improved our ⌘K menu, so now you can control Devin from your keyboard. Find, start, and pin ...

Cognition (@cognition) — Wed, 26 Aug 2026 16:00:22 GMT

We also improved our ⌘K menu, so now you can control Devin from your keyboard. Find, start, and pin sessions, navigate the app, and more.

Your browser does not support the video tag.

💬1🔄1❤️18👀2703📊3

The Devin webapp is getting a facelift. 80% less loading lag, customizable sidebar, do everything wi...

Cognition (@cognition) — Wed, 26 Aug 2026 16:00:20 GMT

The Devin webapp is getting a facelift. 80% less loading lag, customizable sidebar, do everything with ⌘K, and more.

Your browser does not support the video tag.

💬20🔄22❤️309👀145389📊62

@rohlikgroup Read all about it: https://t.co/urD4iqko2G

Cognition (@cognition) — Tue, 25 Aug 2026 16:54:35 GMT

@rohlikgroup Read all about it: devin.ai/customers/rohl…

💬0🔄0❤️41👀10946📊6

In Prague, Czechia, you can order a week's worth of farm-fresh groceries and have it at your door in...

Cognition (@cognition) — Tue, 25 Aug 2026 16:39:32 GMT

In Prague, Czechia, you can order a week's worth of farm-fresh groceries and have it at your door in an exact 15-minute window.

Milk in glass bottles from a local family farm. Fish caught yesterday. Strawberries picked that morning.

This is all fueled by @rohlikgroup, a powerhouse from Czechia. 

It’s profitable, it did >$1.3B of revenue last year, and it's accelerating European grocery delivery with Devin.

Your browser does not support the video tag.

💬33🔄67❤️991👀236719📊199

Try it out today: https://t.co/oeNpdLaq5e

Cognition (@cognition) — Fri, 21 Aug 2026 21:08:27 GMT

Try it out today: 

devin.ai/download

💬1🔄1❤️23👀4978📊3

GPT-5.6 Sol is now the most affordable frontier model on Devin Desktop and CLI. Today, OpenAI reduc...

Cognition (@cognition) — Fri, 21 Aug 2026 21:08:26 GMT

GPT-5.6 Sol is now the most affordable frontier model on Devin Desktop and CLI.

Today, OpenAI reduced Sol’s prices by 20% for the next 3 months. Last week's 70% discount still applies, so through October 3 you now get a 76% discount off the list price.

💬38🔄43❤️935👀100912📊142

@SlackHQ Proud to be a launch partner with @SlackHQ. https://t.co/9AB5hlptTW

Cognition (@cognition) — Thu, 20 Aug 2026 18:55:45 GMT

@SlackHQ Proud to be a launch partner with @SlackHQ.

x.com/SlackHQ/status…

Slack@SlackHQ

x.com/i/article/2090…

💬0🔄0❤️26👀6658📊2

Introducing @SlackHQ Code in Devin. Now, Devin proactively creates dedicated code channels to keep ...

Cognition (@cognition) — Thu, 20 Aug 2026 18:40:46 GMT

Introducing @SlackHQ Code in Devin.

Now, Devin proactively creates dedicated code channels to keep work focused.

Read more about how we tuned Devin’s personality in Slack:

devin.ai/blog/devins-sl…

💬14🔄34❤️278👀44590📊56

It's time to be ambitious.

Cognition (@cognition) — Wed, 19 Aug 2026 18:44:49 GMT

It's time to be ambitious.

Scott Wu@ScottWu46

At Cognition, one of our values is to go for it all: when faced with a tradeoff, pick the ambition-maximizing direction.

When we launched Devin in 2024, we envisioned a future where every team had an infinite army of junior engineers. We were early and Devin wasn't good enough.

Now it is! The only constraint left is how ambitious you're willing to be.

💬9🔄6❤️183👀34687📊21

Thanks for the great week, Tashkent! #ioi2026 Cognition is proud to sponsor this year's IOI and sup...

Cognition (@cognition) — Tue, 18 Aug 2026 20:28:39 GMT

Thanks for the great week, Tashkent! #ioi2026

Cognition is proud to sponsor this year's IOI and support the new generation of competitive programmers.

💬10🔄11❤️235👀45348📊34

GE built America's first jet engine in 1942. Today GE Aerospace's engine technology powers 3 of 4 co...

Cognition (@cognition) — Mon, 17 Aug 2026 15:47:10 GMT

GE built America's first jet engine in 1942. Today GE Aerospace's engine technology powers 3 of 4 commercial flights in the world.

Now they’re using Devin to build more, faster. One team nearly doubled its engineering output.

Their software optimizes routes, reduces fuel consumption, and monitors anomalies.

Their software team's goal is to fly more without adding another airplane, runway, or pilot.

Your browser does not support the video tag.

💬26🔄87❤️996👀358886📊179

The worst interactions are: 3. Not interested (-43.2) 2. Mute author (-58.8) 1. Report post (-234.0...

Cognition (@cognition) — Thu, 13 Aug 2026 20:20:53 GMT

The worst interactions are:

Not interested (-43.2)
Mute author (-58.8)
Report post (-234.0)

💬3🔄5❤️151👀14774📊22

Explore the full repo with DeepWiki, the free codebase understanding tool by Devin: https://t.co/Wlz...

Cognition (@cognition) — Thu, 13 Aug 2026 20:20:53 GMT

Explore the full repo with DeepWiki, the free codebase understanding tool by Devin:
deepwiki.com/xai-org/x-algo…

💬0🔄5❤️151👀20152📊36

The weights for the X algorithm have been publicly released. The highest boost to a post comes fro...

Cognition (@cognition) — Thu, 13 Aug 2026 20:20:52 GMT

The weights for the X algorithm have been publicly released. 

The highest boost to a post comes from copying its link to share it. (Try it on this one for a surprise!)

💬51🔄150❤️3077👀160924📊616

Gemini 3.7 Flash is now available in Devin. On FrontierCode 1.1, it reaches Claude Sonnet 5-level p...

Cognition (@cognition) — Thu, 13 Aug 2026 20:08:46 GMT

Gemini 3.7 Flash is now available in Devin.

On FrontierCode 1.1, it reaches Claude Sonnet 5-level performance at less than half the cost, with the low latency the Flash series is known for.

💬25🔄32❤️469👀153337📊80

Within Devin, Gemini 3.7 Flash performs particularly well on tightly scoped refactors, where it deli...

Cognition (@cognition) — Thu, 13 Aug 2026 20:08:46 GMT

Within Devin, Gemini 3.7 Flash performs particularly well on tightly scoped refactors, where it delivers minimal diffs that match repo conventions.

Gemini 3.7 Flash is available at a 50% discount for 2 weeks, through 08/27/2026.

Read more: devin.ai/blog/gemini-37…

💬6🔄10❤️156👀60911📊22