Latest AI models, one subscription
Picking a model used to be a side job. One subscription for chat, an API key for the new release everyone’s talking about, a second key because the first provider was down, a spreadsheet to remember which one was good at code. You were doing procurement when you wanted to be doing work.
Skynet collapses that. Every frontier model lives in the same workspace, on the same subscription, behind the same prompt box. When a new model ships, it shows up in your catalog — you don’t chase it, file a request, or wait for a renewal. The thing you were going to read a launch thread about is already there to try.
Let routing decide, or pick your own
The default is smart routing. You write a prompt; Skynet reads it and chooses the model that fits. A quick reformat goes to something fast and cheap. A gnarly refactor or a long reasoning chain goes to a frontier model that earns its keep. You get a good answer without auditioning four models yourself.
When you’d rather drive, pin a model. Want this whole conversation on Claude Opus? Pin it. Curious whether Gemini handles a long document better? Switch to it for the next message — same thread, same context — and keep going. Routing is a default, not a cage.
Both modes run on the same credits. Every model has a price, and a heavier model costs more per task than a light one — which is exactly why routing is worth having. You see the cost of a task as you run it, so “use the best model” stops being a blank check. There’s a free tier to feel it out before any of that matters.
A typical first hour looks like this:
Open Chat and just type
Start in Chat and write your prompt like you’d write it to a sharp colleague. Don’t pick a model. Routing makes the call and you read the answer.
Pin a model when it matters
Doing something where the model choice is the whole point — a tricky proof, a tone-sensitive rewrite? Open the model picker and pin a specific one for the conversation.
Switch mid-thread to compare
Not happy, or just curious? Swap models on your next message. The thread and its context carry over, so you’re comparing answers on the same problem, not starting from scratch.
Watch the credits
Each reply shows what it cost. Over a few tasks you’ll learn where a frontier model pays off and where a lighter one is plenty — and routing already knows that, so you can hand the wheel back.
What the catalog gives you
Browse everything under Models: the full lineup, what each is good at, and the routing rules that decide between them. It’s also where you tune the parts that are yours to tune.
- Smart routing — one prompt, the right model, balanced for quality, speed, and cost.
- Pinning and mid-chat switching — override the default whenever you want, without losing context.
- New releases on arrival — frontier models land in your catalog as they ship, no extra subscription.
- Fine-tuning — adapt a model to your data when off-the-shelf isn’t close enough.
- Credit-based pricing — transparent per-task cost, with a free tier to start.
The point isn’t that Skynet has a lot of models. It’s that you stop thinking about which one. You describe the work; the right intelligence shows up to do it.
Frequently asked questions
No extra subscription. New frontier models appear in your catalog as they release, and you use them on the same credits as everything else. Heavier models spend more credits per task than lighter ones, and you see that cost as you run it.
Yes. Smart routing is the default, not a requirement. Pin any model for a conversation from the model picker, and you can switch to a different one on your next message whenever you want to.
It carries over. The thread and its context stay intact, so the new model picks up where the last one left off. That is what makes mid-chat switching useful for comparing answers on the same problem.
Pricing is credit-based. Every model has a per-task price, so you can see exactly what a task costs as you go. There is a free tier, so you can try routing and a few models before committing.