Priyanshu Verma

20 | Engineer | Vibe coder

[Hashnode @hashnode]

LLM rate cards hide the two rules that actually move invoices.

Anthropic's models from 4.7 onward tokenize about 30% heavier. Same listed price, ~30% bigger bill, invisible unless you log token counts.

And long context: OpenAI and Google surcharge it (2x input, 1.5x output past a threshold), Anthropic doesn't. GPT-5.6 Sol is 9% pricier than Opus 5 on a normal coding request. Feed both a 300K-token repo and Sol costs $3.09 against Opus 5's $1.55.

The vendor rankings flip depending on your workload shape. Verified rates for 12 models: https://t.co/wNvY6PkQhw

[Bug0 @bug0inc]

QA has reached new heights.

[Priyanshu Verma @PriyanshuV_code]

finally claude is clauding

[Priyanshu Verma @PriyanshuV_code]

@ay_ushr @text_chorus 😂😂

[Priyanshu Verma @PriyanshuV_code]

@Im_IrushiK its a set of random characters that tells your ownership over something!!!

[Priyanshu Verma @PriyanshuV_code]

more 50% X Algo written in rust

[Priyanshu Verma @PriyanshuV_code]

@BytesWithAnkit thats why i always say "hi" first

[Priyanshu Verma @PriyanshuV_code]

@Harish_521 i think he doesn't like what you just asked

[Sandeep @Sandeepg33k]

🚀 QoL Improvement Alert

[email protected] just made AI step execution way more flexible.

You can now seamlessly combine snapshot-based execution with CUA (Computer Use Agent) and control it at multiple levels:

- Globally
- Per runStep call
- Or even per individual step

💡 This gives you precision where you need it, and speed where you don’t.

You can also contribute and join us:

https://t.co/8CUR3RCT0H

[Priyanshu Verma @PriyanshuV_code]

@Amank1412 😂😂

[Sandeep @Sandeepg33k]

🎉 [email protected] now supports CUA mode!

Until now, passmark drove the browser through ARIA accessibility snapshots - fast and reliable when the DOM is well-structured, but it can struggle with canvas-heavy apps, custom elements, or markup that doesn't expose a clean a11y tree.

After working with a lot of customers, we realized that the winning combo is ARIA snapshot + CUA.

In CUA mode, the agent sees screenshots, clicks coordinates, and types like a human - powered by the new @OpenAI gpt-5.5 computer-use model.

Snapshot mode is still the default and what we recommend for most flows. CUA is the fallback for when the accessibility tree isn't enough.

One flag to switch:

configure({ ai: { mode: "cua" } });

Requires an OPENAI_API_KEY. Would love feedback from anyone testing flows that snapshot-based agents have trouble with.

Passmark is OSS.

Check it out and contribute:
https://t.co/8CUR3RCT0H

Shout-out to @IpseetaP for the PR!

[Priyanshu Verma @PriyanshuV_code]

Is weekend ka toh ho gaya

[Priyanshu Verma @PriyanshuV_code]

i'll be adding lizard to eat the bugs

[Visual Studio Code @code]

Add a little fun to your coding sessions 🐶🐱
This extension lets you keep a virtual pet right inside VS Code.

[Priyanshu Verma @PriyanshuV_code]

@BytesWithAnkit 4 years ago my friend told me that his future is dark,
now i know what did he mean

[Priyanshu Verma @PriyanshuV_code]

@BytesWithAnkit AGI would have been achieved then, i guess

[Fazle Rahman @fazlerocks]

passmark is so close to 500 stars…

check out the most stable and reliable open-source AI regression testing library built on Playwright by the Bug0 team

https://t.co/SVhZHNopo6

[Fazle Rahman @fazlerocks]

Shipping Bug0 Browsers today

Real Chromium in the cloud. One API call, one CDP URL. Your Playwright code runs as-is.

Sessions spin up in seconds and shut down when you're done. Built for speed. One thing done well.

Your agents, crawlers, and test suites need it.

[Sandeep @Sandeepg33k]

[email protected] supports @Cloudflare AI Gateway!

btw, we just crossed 200 stars! 🎊

We're just getting started. If you are an SDET you can use Passmark for free within your org and let AI upkeep the burden. Check it out and feel free to contribute.

https://t.co/VNoXQZbm7G

[Priyanshu Verma @PriyanshuV_code]

i can feel it

[Priyanshu Verma @PriyanshuV_code]

@BytesWithAnkit i am sending this to my cousins under 18