🆕 @OpenAIDevs GPT-5.5 is now generally available and rolling out in GitHub Copilot. Our early testing shows ➡️ It delivers its strongest performance on complex agentic coding tasks ➡️ It resolves real-world coding challenges previous GPT models couldn’t Try it out in Copilot CLI or @code. 👇 github.blog/changelog/2026…

@github @OpenAIDevs OpenAI increased the price 2x, and token usage is supposedly better ...and you jack the price 7.5x ?????

@github @OpenAIDevs x.com/MithunShanbhag…

@github @OpenAIDevs Whelp, time to move on.

Grok Imagine now has dramatically improved lip sync and sharper audio quality on all image-to-video generations. Dialogue tracks the mouth. Sound matches the scene. Your videos look and sound the way you imagined them. Try it today in the Grok app.

@github @OpenAIDevs Oh, wow a 7.5x multiplier, makes sense as it's more expensive compared 5.4, but still! Also Pro+ only, I used to recommend people Copilot to test out the frontier models at a fraction of cost, not anymore!

@github @OpenAIDevs Microsoft said Copilot drove over 40% of GitHub’s FY2024 revenue growth, with GitHub now at a $2B revenue run rate. So making GPT-5.5 a 7.5x premium request multiplier feels ridiculous. Make it 1x or 2x. microsoft.com/en-us/investor…

@github @OpenAIDevs A really odd time to massively hike pricing just when copilot seemed to be gaining momentum.

@github @OpenAIDevs Copilot just went from helpful intern to senior engineer overnight. Agentic mode is actually shipping features now.

@github @OpenAIDevs 7.5x pricing? you guys REALLY want copilot to go downhill, don't you?

@github @OpenAIDevs Fix the rate limit s*** first! Before can even think of going back to GitHub copilot!

@github @OpenAIDevs I am pretty sure that microsoft knows they are pushing users away from copilot. 7.5x and they are also planning on moving to token based usage. While openai only increased the price by 2x, with better token usage.

@github @OpenAIDevs I'll give it a go. As per openai it uses less tokens so the 7.5x shouldt be a big deal. Otherwise, sonnet + 5.4 is the sweet spot

This is where model upgrades start to feel real, not just benchmark-real. Once agentic coding tasks get meaningfully better, product expectations around developer workflows change fast. Curious what people here think gets rebuilt first: IDE UX, copilots, or the workflow layer around them?

Accelerate deployment with Supermicro Global Services. From data center design and solution validation to on-site deployment and ongoing support, integrated services help reduce time-to-online, improve installation quality, and support mission-critical uptime.

@github @OpenAIDevs Not going to touch it.. that thing gets your account week ban rate limited in minutes. People reported having multi day bans on Pro+ licenses before even the 2nd prompt was finished. Copilot is not a platform you can trust anymore

@github @OpenAIDevs "Complex agentic coding tasks" sounds great until you realize Copilot is still a single-agent tool. The bottleneck in production isn't the model - it's constraint files, recovery logic, state management. Ship GPT-10, won't matter without multi-agent coordination.

@github @OpenAIDevs Absolutely ridiculous rug pull on GitHub copilot pro users who grew accustomed to the best frontier models in vscode

@github @OpenAIDevs 5.5 cost 7.5x or 15x is fine. But can you share with us whether you would keep 5.4 forever at 1x? We just want some level of certainty.

@github @OpenAIDevs Copilot now has hidden session and weekly caps, meaning you’re restricting both token usage and the number of calls. This is too mean.

@github @OpenAIDevs It's so bizarre... 2x the API price but much more token efficient compared to 5.4... why a 10x price bump? (7.5x is the promotional pricing)

@github @OpenAIDevs GPT-5.5 in Copilot is a massive upgrade for agentic workflows. The jump on complex multi-step tasks is real — just tested it on a legacy codebase refactor and it handled context across files much better than before. Copilot is genuinely becoming the best AI coding environment.

@github @OpenAIDevs interesting timing right after the claude code postmortem. real signal here isn't 'gpt-5.5 wins copilot'. it's that enterprise wants swappable model backends. nobody wants to lock into one provider for coding agents anymore. github should support both natively

Starlink’s high-speed internet is available in your area. Experience speeds up to 400+ Mbps to stream your favorite shows and sports, work from home, browse social media and more.

@github @OpenAIDevs Fix the rate limits bs - I’m increasing my budget so I can keep using the service way beyond the standard limits, and you’re still restricting me with limits? This feels like a belated April Fools’ joke.

@github @OpenAIDevs If only it was available to actually sign up for the paid plans right now to USE it

@github @OpenAIDevs Sorry @burkeholland but Fledgling instantly made me remember of Terry's game

@github @OpenAIDevs Curious what “agentic coding tasks” means in your eval. Issue triage, multi-file edits, and CLI workflows are where models usually fall apart.

@github @OpenAIDevs I’d like to use the Copilot’s interface with my Claude Code license, but I’d must pay both Copilot and Anthropic so I use Claude directly.

@github @OpenAIDevs that 7.5x hurts though. when 5.4 is just 1x.

@github @OpenAIDevs genuinely curious how it handles agent memory between sessions. that's where my typescript agents always break down - context gets weird after a few back and forths

@github @OpenAIDevs Give me back the individual Copilot Pro Version. I accidentally canceled it 😐 I will search for another subscription otherwise!

@github @OpenAIDevs But not on Pro. So the right move was to cancel for Codex, since the only two models GPT5.5 and Opus are not available on pro

@github @OpenAIDevs can't wait to solve another real wold coding crud challenge with another SOTA model. super excited

The average American has about $5,700 in gold jewelry sitting around their home. Unvault has created a streamlined way to understand what you own, then you can choose to safely sell the gold. This could be incredibly impactful for a family.

@github @OpenAIDevs Be careful out there. ⚠️ GitHub merged PRs incorrectly yesterday (4/23) due to a merge queue regression affecting squash & rebase merges. This is an Industry standard platform, reputation on the line githubstatus.com/incidents/zsg1…

@github @OpenAIDevs 7.5x usage feels really high especially when your entire usage count is per-prompt.

@github @OpenAIDevs 7.5x

@github @OpenAIDevs 7.5x is oof

@github @OpenAIDevs Another 7.5x model holy shit rate limited must get highered then cause 7.5 x is huge for two new models

@github @OpenAIDevs has to be a fucking joke, 7.5 again. who tf you think you are