messimaniac (1 last dance🐐)

@shregupta89 @kipupwidanshika Yeah Zomato provides convenience..but to save money one can call

@mexavin @kipupwidanshika But that was the whole point of zomato right That we wouldnt have to search every restaurants number separately Plus a lot of them do not have their own websites or number to order from

"Necessity is the mother of invention " Zomato would really have to change their prices now

@shregupta89 @kipupwidanshika So u place the order yourself over phone

@kipupwidanshika But all delivery guys wouldnt place an order and wait that much for u

@OnlyTerp @cjzafir How many dollars of usage is included..trying to compare it with codex sub

@cjzafir I get like infinite deep-seek usage with my ollama cloud membership

Codex for planning Deepseek v4 for execution Trust me!

@cjzafir How about 5.5 for plan, 5.3 codex for execution...

@lexrus Better to plan with 5.5, implement with 5.3 codex or deepseek and review with 5.5 again

I tried having DeepSeek V4 Pro write the implementation plan, then asked GPT-5.5 to review it. It found problems everywhere and basically nuked the whole thing and rewrote it from scratch.

@iwanttobeinmars How is it as compared to codex

ampcode is goated, holy cow haha

@teobucos How is it compared to codex

Also, credit where it’s due: from my own experience, Amp is by far the #1 coding harness out there. Kudos to the team 👏🏼

$819 later, @ampcode discovered I treat my AI usage like a challenge 😆 I'm trying to tokenmaxx all my subs and usage every day. I knew the $15 a day wasn’t going to be around for long after they stopped ads, so I tried to make the most of it. Completely understand and thank you for the very generous run 🙏

@sladurb @sqs And Chatgpt sub will incur extra cost...will that cost be used from the 5$?

@sqs code fre got cut down from $10/day to $5/day. Was that announced beforehand anywhere?

@shikhr_ @thsottiaux True... compaction works well

@thsottiaux I don't even have to change harness because codex is so good

A little secret. About 5% of our production traffic is on the Pi harness, about another 5% is on OpenCode. Reminder you can use your ChatGPT account in a flourishing set of other tools. We’ll continue to make Codex awesome, but you have options.

@0xSero Windows?

DeepSeek-V4-Pro & Kimi-K2.6 running in Codex app. Cheapest way to taste the frontier. Works w local models and they can do all the computer use stuff (:

@thsottiaux Feature parity in vs code, cli and app

For those of you living inside the codex app, what should we prioritize among features, reliability or performance?

@shikhr_ I wish it worked with VS code extension

HOLY SHIT! I GOT CODEX MOBILE APP WORKING ON LINUX!!! I patched the Linux Port of Codex Desktop App Using Codex to enable the hidden feature which connects to Mobile APP. AND IT FRICKING WORKS!!! GPT 5.5 IS AGI

@OpenAI When in vs code extension?

Rolling out today as a preview on iOS and Android in all supported regions. Support for connecting your phone to the Codex app on Windows is coming soon. openai.com/index/work-wit…

You've been asking for this one... Now in preview: Codex in the ChatGPT mobile app. Start new work, review outputs, steer execution, and approve next steps, all from the ChatGPT mobile app. Codex will keep running on your laptop, Mac mini, or devbox.

@steipete Wish it was in Windows

Codex was debugging a Telegram issue and needed a new token, so it used Peekaboo to open the Telegram Mac app, talked to botfather and just did it. Computer Use is amazing. peekaboo.sh

@ryan0x44 @sqs @OpenAI @AmpCode Not open for all

@mexavin @sqs @OpenAI @AmpCode That's what I'm doing.

I've been using @OpenAI Codex today and @AmpCode extensively, both using GPT 5.5 under the hood. First, GPT 5.5 is an unbelievably good coding model. Second, the speed and improved context when using Amp proves the harness can create a shockingly large advantage. Both great tho!

@cjzafir I feel even for planning 5.5 medium is good enough. I tried gemini 3.1 pro and deepseek 4 pro on a bug for 3 days...codex solved in 1 try

How Codex 5.5 medium is beating Codex 5.5 extra high? Add this rule in Agents. md: "Don’t fight errors! Whenever you encounter the same error twice, research the web and find 3-5 possible ways to fix it. Then choose the most efficient solution and implement it." This prompt release stress to think hard and run into a loop of bad decisions but find new perspectives that actually worked. I ran this experiment on complex tasks (model training, running evals, converting best checkpoints to GGUF etc) and I encountered NO ERROR. Best part: minimal 5 hr and weekly limit usage. I am using Codex 5.5 medium (fast) and it feels instant. But here's the catch: Your session plan has to be designed by Codex 5.5 extra high! Use the medium variant, only to execute the plan. This way you're not using Codex 5.5 medium's loose thinking but using its fast speed to execute the task list properly. Try it yourself.

@shikhr_ I tried OC go for 3 days for an implementation plan made by gpt 5.5. DSV4 flash always stuck at a point. Gemini flash same..Then I tried with gpt5.5 medium. Same prompt, Consumed my 5 hr limit but found the bug. So I feel OCgo is best for tasks requiring minimal thinking.

Might as well use OpenCode Go since Codex limits are dead right now. Which models work best?

@hiarun02 Which tier

One five hour session is 25% of weekly quota now.

@sqs I wish you offered something like Opencode go

We have a lot more coming, but we wanted to ship this now because it's such a big improvement. Happy coding!