In brief: The lowest published rate is usually not the lowest bill. Real cost depends on your input-to-output token ratio, whether caching applies, which access tier your usage unlocks, and which costs sit outside per-token pricing. Compare on total cost for your own traffic mix.
Input and output are priced differently
Most models charge more for output tokens than input tokens. If your workload is long input and short output — summarization, classification, extraction — input pricing dominates. If it is short input and long output — generation, writing — output pricing dominates.
So before comparing providers at all, estimate your own input-to-output ratio.
Cache pricing changes the math
If the same prefix is sent repeatedly — a long system prompt, a fixed context block — providers that support prompt caching can reduce cost substantially. Cache policies differ between providers, so check each one's documentation rather than assuming parity.
Access tiers, not just unit price
Many gateways organize access by cumulative top-up amount: lower tiers reach a subset of models, higher tiers unlock more. That means a low unit price can come with a narrower model selection.
FlowBarAI works this way: a free trial tier requires no top-up, and cumulative successful top-ups at $10, $30 and $80 unlock progressively broader model groups. The specific model list changes over time, so check the current model catalog.
What free actually means
People searching for a free LLM API usually want to prove an integration works before committing. The questions that matter are: is the free allowance one-time or recurring, does it require a card, and are the free models available long term or only during a trial?
A cost comparison method you can reuse
Write down your real input and output token volumes. Compute input rate times input volume plus output rate times output volume. Confirm which access tier your usage reaches. Check whether caching or batch processing applies. Only then compare totals across providers.
Frequently asked questions
What is the cheapest LLM API?
It depends on your input-to-output ratio, whether caching applies, and which access tier your usage unlocks. Compare on total cost for your actual traffic mix rather than on a single headline rate.
Are free LLM APIs really free?
Free allowances differ in form: one-time trial credit, recurring quota, or permanently free models. Confirm which one applies, whether a card is required, and whether the free models are available long term.
Does FlowBarAI require a subscription?
No. It is pay-as-you-go based on actual token usage, with cumulative top-up tiers that unlock broader model groups.
How is FlowBarAI billed?
By actual token usage. Current rates and top-up tiers are published on the pricing page, and every paid request is quoted from the live catalog before submission. Failed requests are not charged.
See the current rates
Pay-as-you-go pricing, published rates, and per-request quotes before submission.