Cosine

The coding agent for engineers with taste. Complete complex tasks in your codebase with clarity and control. Available across CLI and Cloud.

[Cosine @CosineAI]

We are proud to announce Cosine's selection for the UK Government's @UKSovereignAI initiative.

As one of the first companies selected, we will train, own, and deploy the first sovereign AI model built entirely on UK soil.

Read more below.

[Rachel @RachBLondon]

If you're in London and building with AI, you should come to this.

We're partnering with @AntlerGlobal for Deploy at Google London on 10 June. You'll get access to @CosineAI coding agent, a room full of ambitious builders, and prizes for the best projects.

Join us!👇
https://t.co/mHj3b6DLt4

[Cosine @CosineAI]

Cosine is an AI coding agent for real software engineering.

It scales from quick fixes to complex migrations, working in your codebase via CLI or Cloud.

Available on macOS, Linux, and Windows.

[Cosine @CosineAI]

Gemini 3.5 Flash is now available in Cosine.

With impressive response times, our engineers have found that the model is best suited for high-frequency code edits, rapid-fire debugging, and swift repository-wide exploration.

[Yang Li @yangli_]

Excited to announce Tom Dowley is our new Head of Engineering at @CosineAI

A rare mix of intellectual curiosity and pragmatic, charismatic leadership. Ready to build the best with him.

[Cosine @CosineAI]

Build an offline AI agent. Join Cosine's track at
@OverMindLab’s Localhost Agentic Hackathon on 20 June.

Design, build, and deploy autonomous agents onto physical hardware with zero internet connectivity.

Hosted by @dawncapital with limited spaces:

https://t.co/sXti0QUWi8

[Rachel @RachBLondon]

From a hand-tracking AR app to a dentist’s revamped landing page — so many cool concepts hacked out with @CosineAI at @RamenClubHQ

[Cosine @CosineAI]

AI coding is now 55% of enterprise inference, but it’s blocked by three core fears: security, cost, and control.

At Cosine, we solved them:

→ Extreme security, even air-gapped
→ Lumen Outpost: Our specialized model for max efficiency and lowest cost
→ Real-time traceability

[Yang Li @yangli_]

AI is reshaping software & services, but execution is challenging.

With the £500M Sovereign AI fund backing UK startups like @CosineAI the UK has a critical window to lead.

Join our private breakfast panel on June 9th in London to discuss the future of UK AI leadership ↓

[Robert Gibson @RobGibson20]

One year ago, I joined @CosineAI as an intern. Back in January, I took a full-time Product Marketing role mid-dissertation.

I went from feeling behind to achieving more than expected in 12 months. Been a great ride so far.

Thanks to the team – looking forward to what's next.

[Cosine @CosineAI]

Cosine is partnering with @AntlerGlobal for their Deploy hackathon at Google London on 10 June.

Build production-grade software with our cutting-edge coding agent. Cash prizes.

Hosted by @GoogleStartupUK and credits by @Lovable – with tracks from @OvermindLab, @paid_ai, Ralio, + Listo.

Limited spaces. Grab yours: https://t.co/HXLtavjorz

[Cosine @CosineAI]

Niche language codebases deserve better support.

Lumen Outpost – our latest coding model built for engineers – fixed broken tests in a niche Elixir codebase and got the full test suite passing in under a minute.

Try now with Cosine CLI: brew install CosineAI/tap/cos

[Rachel @RachBLondon]

Hackday at @RamenClubHQ next week👷🛠️

Builders, side projects, demos, and access to the latest Cosine agent for anyone who wants to ship something cool.

Dalston, London 👇
https://t.co/CcMhZA0CDM

[Cosine @CosineAI]

Lumen Outpost benchmarks reveal how targeted post-training transforms model capability, rivaling frontier models.

Highlights:

→ Niche-Bench: Leads with 53.9% pass rate (e.g., +11pp ABAP)
→ Slop-Bench: Highest quality score (25.4%)
→ Vibe-Bench: Near-perfect agent alignment (96.3% Action)
→ Cost-Efficiency: Strongest profile at $7.90 per successful task

Read the full report: https://t.co/3RfZnspYo0

[Alistair @AlistairPullen]

This was a great first step for us in terms of focusing on a specific set of problems within agentic coding - performance across prolific programming languages and amount of sloppy code the model writes we focused on both of these during our RL post-training phase

[Cosine @CosineAI]

Lumen Outpost is available now via API and in Cosine.

These results show our post-training stack can push models toward the frontier, forming the blueprint for our upcoming Lumen Sovereign work.

Full release:
https://t.co/0HiFGFUUPh

[Cosine @CosineAI]

Introducing Lumen Outpost.

AI models usually try to be everything to everyone. We did the opposite. Every training decision we made for our new model was centered on a single objective: writing exceptional code.

Outperforming frontier models. Available now.

[Cosine @CosineAI]

Lumen Outpost delivers exceptional results across our key benchmarks.

Outpost is 3x more cost-efficient than GPT-5.5, while also outperforming frontier models in niche languages, generating more maintainable code, and aligning with our preferred behavioral vibe signals.

[Rachel @RachBLondon]

I’m such a diagram nerd. Watching a swarm of agents visualise itself in real time is very satisfying!