Five real schoolwork-shaped tasks (an essay, a summary, a grammar check, a to-do list, a translation) priced per model with this site's actual cost formula (real token counts × each model's real price per token). Never the globally cheapest model if this site's own fit judgment grades it a bad fit for the task.
These are API costs, billed per token, the same number every other page on this site uses. They're not the same as what you'd pay through a chat app: some models below have a genuinely free, rate-limited API tier; some chat apps (including free ones) offer flat monthly subscriptions instead of per-token billing. API cost is what this site can measure the same way for every model, so it's what's shown here.
"Good fit" means this site's own recommendation logic graded that model green for this task's difficulty: the same judgment shown under every race result. See how we measure →.
Total cost to run all five tasks once on each model, ranked cheapest first:
The cheapest model here has a genuinely free API tier (rate-limited, not a placeholder): $0.00 for all 5 tasks. The most expensive model on this list would still cost $0.39 for the exact same five tasks.
Because cheapest and "good enough" aren't always the same model. A model too small for a task can come back wrong or thin; cheap and wrong is still wrong. Every price above is paired with this site's real fit judgment, not cost alone, so a student uses the cheapest model that can actually do the assignment, not just the cheapest model, period.
Every figure above is a real computed cost: same formula as every result on this site, never a marketing number. Want a plain-language pick for your own task? Which model should I use? → Or see the full cost comparison → across all 25 tasks.