You can use Solar Pro 4 before paying anything, and when you do pay, the two things worth understanding are the token rates for API work and the monthly plans that bundle credits for chat and agent sessions.
Start free
New accounts start with free credits after signing in with Google on the home page. That is enough to put the model on a real task — a long document set, a terminal job, a multi-turn tool call — before you decide anything about a plan.
API token pricing
| Item | Price |
|---|---|
| Input | $0.30 / 1M tokens |
| Cached input | $0.06 / 1M tokens |
| Output | $1.20 / 1M tokens |
A rough example: an agent run that reads 100k input tokens and returns 8k output tokens costs about $0.04 — $0.03 for the input and roughly $0.01 for the output. Repeated calls over the same documents hit the cached-input rate, which is five times cheaper than fresh input.
Plans
| Plan | Price | Credits | Made for |
|---|---|---|---|
| Starter | $19 / month | 60 credits | Exploring agentic conversations and lighter document work |
| Pro | $39 / month | 150 credits | Teams running multi-step jobs across documents and tools |
| Premium | $79 / month | 400 credits | Continuous agent workloads, concurrent sessions, history export |
Every plan includes access to Solar Pro 4 chat and 512K context sessions; Pro and above add priority queues and the full 128K output ceiling for long deliverables. Current details and checkout live on the pricing page.
Which one do you need?
- Free credits first. They are the cheapest way to find out whether the model finishes your task.
- Starter if your work is occasional: a few long documents, some tool calls, no queue pressure.
- Pro if agents run every day and output length matters.
- Premium if sessions overlap or you need history and export across a team.
FAQ
Is there a free tier?
New accounts start with free credits; no payment details are needed to start chatting. See the home page to sign in.
How is API usage billed compared with plans?
API usage is billed per million tokens at the rates above. Plans bundle monthly credits for chat and agent sessions — compare them on the pricing page.
What makes cached input cheaper?
Cached input is charged at $0.06 per 1M tokens, so repeated calls over the same documents cost a fraction of fresh input. Long-context agent loops benefit the most.
Where are the technical limits documented?
The API guide covers context, output limits, and the OpenAI-compatible setup.
