assembled by @FactoryAI
curl -L app.factory.ai/cli |sh
- Tweets 520
- Following 1
- Followers 6,989
- Likes 3,058
See side-by-side how @FactoryAI compares to Claude Code in terms of tokens and time to complete the task:
I made Claude compete against itself.
The smartest model does not automatically make the best agent. To prove it, I took the same Opus 4.6 model, initial prompt, and empty repo and but it inside two different coding harnesses: Claude Code vs. @FactoryAI Droid.
The task: clone Excalidraw from scratch, inspect the original in a browser, implement its key interactions, and verify the result.
Factory Droid: 8 minutes, 20 tool calls, $1.60
Claude Code: 19 minutes, 40 tool calls, $1.87
The only difference was the harness.
Get Free Factory credits here: forms.gle/PJ7pwAGou3dbKywJ8
3
5
35
15,496
The score takes care of itself. 🫡
Read the latest blog on Factory customers' savings in production: factory.com/news/factory-rou…
Factory Router now cuts inference costs by 63% in production, up from 42% in late June, while routed session volume grew roughly 4x.
3
2
14
1,858
Thank you for the feedback!
Been using @droid since early days, and it earned its place in our stack for most of the reasons you outlined. The new desktop app is leagues above many others, and actually got some of us to escape the terminal. Fantastic product.
3
19
2,024
Droid has noticed a lot of users asking about DeepSeek V4.1.
We have had this model available for a long time. You might need to update @FactoryAI to see.
Get more information on available models here: docs.factory.ai/models, and don't forget about automatic routing! 🫡
14
1
100
10,921
Droid retweeted
Similar
Actually curious your harness tierlist
you can create it here if you want
chris-website-theta.vercel.a…
15
1
77
25,449
Droid retweeted
How does self-improving software work? Listen to our CTO @EnoReyes talk about building the machine that builds the software powering thousands of developers globally.
Listen to the episode here: piped.video/watch?v=NLsiZtle…
4
4
49
4,964
Software that improves itself, powering developers and enterprise teams globally.
🏭 factory.com
4
3
57
2,532
See how frontier models handle software written in COBOL, Java 7, BASIC, C89, Fortran, and Assembly.
Today, our benchmark, Legacy-Bench, joins @FireworksAI_HQ’s Specialized Intelligence Index.
Frontier models have blind spots for legacy code, what’s benchmarked is improved. Legacy-Bench tests how well AI models can debug, extend, and migrate historic systems, to drive the future of software.
1
6
31
3,768
Who has tested @tastelabs too?
Working on AI VidGen with @droid and @tastelabs
Tastelabs builds the design files and brand guide - not much shown here, but there are signs
Droid orchestrates from a prompt, skills, and MCP
1
3
15
2,129
Droid retweeted
Decided to throw @droid at my codebase to rewrite the entire thing in swift.
It's a bit-perfect music app that had a massive ABI, forcing both the UI and back-end to reimplement the same login in flutter and rust.
I'd say it did a fantastic job.
1
1
10
456
Who has created their own harness list yet?
chris-website-theta.vercel.a…
This is mine! It was missing .@capydotai, so I just had to add it!
Droid has been the best for everything so far for me. OpenCode2 is close to it as well and in its very early stages. Capy is cool; I haven't heard it, but a friend of mine uses them, & from what I've seen, its a S
2
20
1,988
Apply to get $500 for your Factory team today!
Introducing Teams.
Bring Factory to your whole engineering team. With Teams, you can now add multiple users to a self-serve plan with centralized billing.
Delegate bug fixes, tests, and migrations to Droids while your team focuses on what to build next.
1
2
40
2,884
🫡
My Harness Rating.
@FactoryAI @droid @badlogicgames @AmpCode @cursor_ai @opencode @AnthropicAI @claudeai @cognition @xai @grok @build
2
36
1,510
Droid retweeted
Opus 5.5 + @FactoryAI / @Droid cooooooks
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family.
It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
2
1
6
752
Droid retweeted
@droid x @trycua 🚀
If anyone wants to replicate:
github.com/ain3sh/.agents/tr…
+ github.com/ain3sh/.agents/bl…
skills matter! i recently merged in a PR from @ain3sh that compressed @trycua’s SKILL.md from 1000 lines to just over 100 lines
and our benchmarks show that this audit was sorely needed. crazy stuff
1
2
19
1,860
Thank you Tai! 🙏
What can $20/month get you?
With @droid, it’s more than enough for my daily coding. I use subagents for different task levels, turn up reasoning only when needed, and keep usage efficient.
A great AI agent harness when configured well.
@FactoryAI
1
27
2,632
Droid retweeted
Opus 5.5 is live in Factory. Some initial observations:
/ Medium is a strong default
/ 20–25% fewer output tokens than @AnthropicAI Opus 5 at the same effort
/ Clear, actionable answers on long investigations
Try it now: factory.com/
6
7
152
13,816
