Coding plan
Private by default. US-based model hosting and data storage.
Endless messages for your coding tools (on every plan above $3.99). No message counting. Every request runs at full speed. Big 5-hour windows, a weekly allowance, unused usage banks up to ten days, and a Flash model floor keeps answering from its own allotment if a window fills up. Up to 165M+ tokens a month on top models on the Pro plan and above, and 85M+ on Basic.
Free month of DeepSeek V4.1 Flash 1017, V4.1 Flash, MiMo V2.6 Flash, V4 Flash 0731 and GLM 5.3 Flash. Taster's first month is free: five of everyone's favorite light models. Then $3.99 a month, cancel anytime.
How it works
Everything runs at full speed, always. A 5-hour window starts with your first request and can use up to a fifth of your weekly allowance. Your week resets on the same day every week, and what you don't use banks automatically, up to ten days of reserve. Reach a limit and the response names it and exactly when it lifts, while the Flash models keep answering. Four lines show everything: your window, your week, your bank and your month.
Plans
Concurrency, plainly: per-token accounts run up to 10 concurrent requests and 600 requests a minute. Coding plan concurrency scales with tier: 3 parallel agents on Basic, 5 on Pro, 6 on Elite, 7 on Ultra, 8 on Canopy.
Every plan from Basic up includes every model in the lineup. No model gating: the $12.95 plan calls the same models as the $195 plan, including GLM 5.3, GLM 5.2, GLM 5.3 Flash, DeepSeek V4 Flash 0731, Kimi K2.6, Kimi K3, DeepSeek V4 Pro, MiniMax M3, and Nemotron 3 Ultra. Tiers buy bigger windows, more agents, and a deeper bank.
- Taster, $3.99 a month, first month free: five Flash models (DeepSeek V4.1 Flash 1017, V4.1 Flash, V4 Flash 0731, GLM 5.3 Flash, MiMo V2.6 Flash), 18M+ tokens, API key included.
- Basic, $12.95 a month: 85M+ heavy-model tokens (450M+ on light models), 3 agents at once.
- Pro, $24.95 a month: 165M+ heavy-model tokens (890M+ on light models), 5 agents at once.
- Elite, $50 a month: 335M+ heavy-model tokens (1.8B+ on light models), 6 agents at once.
- Ultra, $99 a month: 665M+ heavy-model tokens (3.3B+ on light models), 7 agents at once.
- Canopy, $195 a month: 1.3B+ heavy-model tokens (6.6B+ on light models), 8 agents at once.
- Up to 30% more usage than the same spend per-token.
- Unused full-speed time banks up for slow weeks.
- A key is included with every paid plan, and so is the full PGS AI app.
- Speaks OpenAI chat completions, Anthropic Messages, OpenAI Responses and the Ollama protocol natively. See native protocols.
Most plans in this industry are built hoping you underuse them. This one is built to be used all the way.
Prefer to pay for exactly what you use? Per-token pricing starts at a $5 minimum with no subscription.