Groq Pricing

Groq pricing is pay-per-use based on tokens consumed; LLM input tokens start at $0.05/million (Llama 3.1 8B) and output tokens at $0.08/million; no subscription tiers required.

Groq is priced per usage rather than per seat. Example: GPT OSS 20B 128k at $0.075 per 1M tokens input / $0.3 per 1M output. Sign up for an API key, then pay per call.

Yes — Groq offers free starter access on signup. Free API key available; pay-as-you-go usage.

Yes — Groq's purchase motion is self-serve. A checkout URL is published on the pricing page. No demo is required.

Yes — Groq's API documentation is publicly accessible, and the API is the primary product surface. An AI agent or developer can inspect the integration surface without authentication.

xpay scored Groq as agent-ready: every advertised plan is self-serve with public pricing. Groq's pricing URL resolves, but plan structure is not surfaced clearly enough for an agent to enumerate without hand-parsing. Per-unit pricing is published — an agent can forecast cost per task before committing. Self-serve signup exists alongside a sales-led / custom tier. Public API documentation is reachable without an auth wall. For the full purchase simulation see the linked Agent-Ready Index scorecard.

In the AI Inference category, xpay's Agent-Ready Index lists alternatives ranked by their ability to support AI agent purchase. Visit the agent-readiness scorecard for Groq to compare with 3 nearest peers.