Key facts
| Free tier | plugsky-micro (reasoning-capable) and plugsky-lite (speed-focused) |
| Balanced paid tier | plugsky-plus — everyday work with tools and vision |
| Workhorse tier | plugsky-pro — general reasoning, code and tool use |
| Top general tier | plugsky-max — the largest general model for hard analysis |
| API | All five use the same OpenAI-compatible chat completions endpoint |
| Context class | 128K-class windows; live limits are published per model |
| Trial | Free plan with no card, plus a 14-day full-access trial |
| Product status | Live |
TL;DR
- Start free: plugsky-micro for reasoning, plugsky-lite for speed.
- plugsky-plus is the everyday paid model with tools and vision.
- plugsky-pro is the production workhorse for mixed workloads.
- plugsky-max is the top general tier — use it as an escalation target.
- All five share one API, so moving up is a model-name change.
How it works, step by step
- Run your real prompts on the free pair, plugsky-micro and plugsky-lite, first.
- Note where the free models fail: reasoning depth, image inputs or long context.
- Move those workloads to plugsky-plus or plugsky-pro and re-run the same evaluation.
- Reserve plugsky-max for the hardest cases and treat it as an escalation target.
- Keep one default and one escalation model per workload in configuration.
- Re-check the live model card whenever the catalogue changes.
Try it yourself
The free pair: plugsky-micro and plugsky-lite
plugsky-micro and plugsky-lite are the two models on the free plan, with no card required. Both support streaming, JSON mode and function calling on the OpenAI-compatible API, and both are 128K-class.
The difference is emphasis. plugsky-micro leans into reasoning, which suits analysis, structured extraction and multi-step chat. plugsky-lite leans into speed, which suits simple chat, classification, tagging and high-volume tool loops. Start every new workload here and only move up when an evaluation says you must.
The paid workhorses: plugsky-plus and plugsky-pro
plugsky-plus is the balanced everyday tier, adding vision and tool use to normal chat flows. plugsky-pro is the general workhorse for production: mixed code, reasoning, extraction and agent steps. For many teams, pro becomes the default paid model once free-tier quality stops being enough.
- Use plus when image inputs or slightly stronger everyday answers are the gap.
- Use pro when the workload is broad and must hold up under production volume.
- Keep one escalation target above whichever model you choose.
Check the live model card for current capability flags before assuming a feature is present on a given tier.
The top tier: plugsky-max
plugsky-max is the largest general model in the catalogue, aimed at complex multi-step reasoning, deep analysis and long-form planning. It is not a default: it is the model you route to when the cheaper tier fails a validator or the request is explicitly high-stakes.
Because all five tiers share the same endpoint, the whole ladder is reachable from one codebase. Move model names into configuration, define a default and an escalation model per workload, and use the model selector to sanity-check choices before benchmarking on your own prompts.
Honest comparison
| Model | Plan | Best for | Escalation path |
|---|---|---|---|
| plugsky-micro | Free | Reasoning-capable free chat, extraction and tool use | plugsky-plus or plugsky-pro |
| plugsky-lite | Free | Fast chat, classification, tagging and tool loops | plugsky-plus or plugsky-pro |
| plugsky-plus | Paid | Everyday work with tools and vision | plugsky-pro or plugsky-max |
| plugsky-pro | Paid | Production workhorse — code, reasoning, agents | plugsky-max |
| plugsky-max | Paid | Hard multi-step reasoning and deep analysis | Human review |
Frequently asked questions
What is the difference between plugsky-micro and plugsky-lite?
Both are free. plugsky-micro leans into reasoning for analysis and multi-step tasks, while plugsky-lite leans into speed for simple chat, classification and high-volume loops. Test both on your own prompts.
Which Plugsky model should I start with?
Start with the free pair, plugsky-micro and plugsky-lite, and only move to plugsky-plus, plugsky-pro or plugsky-max when an evaluation shows a real quality or capability gap.
What does plugsky-plus add over the free models?
plus is the balanced paid tier for everyday work, adding vision and stronger tool use. It sits below plugsky-pro and plugsky-max in the ladder.
Is plugsky-pro good enough for production?
For many teams it is the production default: general reasoning, code and agent workloads with function calling. Keep plugsky-max as the escalation target for hard or high-stakes requests.
When should I use plugsky-max?
For complex multi-step reasoning, deep analysis and long-form planning, or as an escalation target when a cheaper model fails validation. It is the top general tier, not a daily default.
Can I use all five models in one application?
Yes — every tier shares the same OpenAI-compatible endpoint, so choosing a model is a name in configuration. That makes a default-plus-escalation policy straightforward to implement.
Is there a free plan?
Yes. The free plan includes plugsky-micro and plugsky-lite with no card required, and a 14-day full-access trial lets you evaluate the paid tiers.
How is pricing structured?
Self-serve plans are flat monthly with fair-use usage rather than per-token billing. Check the live pricing page for current plans and limits.