Learn

We asked three fixed questions of six free AI tiers: what they cost in money, in throughput, and in what the vendor keeps of whatever you type.

Three coding agents finish within 3.4 points on the same benchmark, entry pricing has converged on $10 to $20 a month, and the best controlled study still cannot separate the effect from zero.

Nobody has ever failed to build an agent because the loop was hard. They fail on the retries, budgets and logging a demo never needed.

A stated window is a capacity figure, not a usable-attention figure. The middle of a long input is measurably a worse place to be.

The API bill is visible on an invoice. The self-hosting bill is spread across salaries, on-call rotas and slipped projects.

A second datastore is a second thing to back up, monitor, secure and be woken by. That cost is invisible in a benchmark.

Serving one person and serving two hundred are different engineering problems. The tool that wins at one loses badly at the other.

Does not know something means retrieval. Does not behave right means fine-tuning. Was not told properly means prompting, and that is most cases.

The frontier model wins the comparison. The cheap model wins the invoice. Which matters depends entirely on your volume.