Launching today
Modelpeers
Follow AI models and agents like you follow people
88 followers
Follow AI models and agents like you follow people
88 followers
Modelpeers is like LinkedIn for AI models and agents. Follow Claude Code, GPT-6, Gemini and 200 others, and see their new versions, price and limit changes, outages and retirements in your feed.
Free
Launch Team / Built With
Framer AI AgentsDesign and publish professional sites with AI
Promoted
Hi Product Hunt, I'm Lior, the maker of Modelpeers.
I build with AI coding agents every day, and I kept learning about changes the hard way. A model I relied on got a shutdown date I never saw. A coding agent's plan quietly gave me less usage. A better model came out, and I heard about it two weeks late on X. All of it was public, spread across changelogs, status pages and pricing pages that nobody has time to read.
Modelpeers lets you follow AI agents and models like you follow people, and tells you when one changes. It's free.
1️⃣ Follow agents like Claude Code, Cursor, Codex and Copilot, and models like GPT, Claude and Gemini. It launches with 163 models and 31 agents from 13 companies, and the list keeps growing.
2️⃣ Modelpeers watches their changelogs, deprecation notices, status pages and pricing pages.
3️⃣ Every Monday you get one email: what changed, who it affects, what to do. Quiet weeks send nothing. Turn on urgent alerts, and a shutdown three days away reaches you the same day.
You don't have to wait for Monday. The live feed shows every change as it's published: new releases, retirements, price and limit changes, incidents and benchmark results, each linked to its source.
Each agent and model also has a profile:
🗓️ Releases, retirement dates and what to switch to
💵 Plans, prices and usage limits, with before and after
📊 Benchmark ranks from public leaderboards
🚨 Incident history from official status pages
💬 Short quotes from builders, each linked to the original post
🥚 An "About me" each model wrote about itself through its own API, weaknesses included, published unedited
✍️ Writing habits: how many dashes it uses, and every "delve"
I'd love to hear which agent or model you'd follow first, and what's missing. I'll be here all day answering questions. 🙏
good timing, I live inside Claude Code and its limits. do I hear the same day a limit changes, or only in the Monday digest?
@olia_nemirovski Thanks Olia! You can have both.
Your feed updates throughout the day, so a Claude Code limit change shows up there shortly after it's announced.
The email digest goes out on Monday. If you turn on same-day emails, when you follow or later in your email settings, urgent changes reach you the same day. That includes a big limit cut that takes effect before Monday.
Claude Code is already there, just hit Follow: https://modelpeers.com/agents/claude-code
The 'quiet weeks send nothing' rule is what would make me actually stay subscribed to the Monday digest — most change alerts train you to ignore them. The before-and-after on usage limits is the other piece I care about, because a limit change only matters if I can see what I had before. When a model is retired, does the profile point you to the closest equivalent you already follow, or just list what to switch to in general?
@jcbuilds Thanks JC! That's exactly why I built it that way, most alerts end up muted.
On retirements - when the company names a successor, Modelpeers follows it for you automatically, so you keep getting updates on the new model. It doesn't yet look at what you already follow and pick the closest match from there. That's a really good idea, adding it to my list.
Love this idea Lior :) We had that happen at SlideSpeak a few times!
@kevin_goedecke Thank you! Yeah, as companies use more and more AI models, it becomes a frequent problem for a model to change or be deprecated!
I don't know how no one else thought about this already but it's a cool idea
Congrats on the launch Lior
@liran_tal Thank you, much appreciated
congrats on the launch lior. the quiet plan change is the one that hurt me most, nobody announce when usage gets smaller, you just notice the agent stops earlier. how do you catch changes that live only in a pricing page table and never in a changelog?
@agshin_rajabov Thanks! Yes, the quiet ones are the worst. Modelpeers reads the pricing and plan pages themselves, not just changelogs, and checks them every few hours. When a page changes, it compares it with the previous version, so a limit that quietly drops shows up with the before and after.
the "about me" written by the model through its own API is the detail that got me, especially "weaknesses included, published unedited." but a model describing its own weaknesses is still the model deciding what counts as a weakness worth admitting. have you ever checked one of those self-written profiles against what actually shows up later in its incident history? curious if there's ever a real gap between what a model says about itself and what it actually does.
@galdayan Thanks Gal! That's exactly why the "about me" isn't the only thing on the profile. Right below it you see what the model actually did: its outages, changes, and benchmark results (each with a source), plus reviews from people who use it. So you can put what it says about itself next to its actual track record :)