Get the most value from OpenCode Go
Every model on the Go $10 plan comes with a
different usage allowance, which makes them hard to compare. So this page lines them all up against
one everyday example: around 100 million tokens of coding work, where most of it reuses things the
model has already seen. Change the numbers below to match how you work — the whole table updates
instantly, all right here in your browser.
How to read: “Cost” is dollars of allowance consumed by the workload above
(input at fresh/cached rates plus output at the full output rate).
“% of $10 cash” = cost ÷ 10 (cash-equivalent). “% of allowance” = cost ÷ that model’s Go allowance
(e.g. $3.73 on a $60 allowance = 6.2%, ≈16.1 blocks per month). Values >100% mean the workload does not fit in one $10 subscription for that model.Assumptions: “Total tokens” splits into input and output by “Output %”;
output tokens are always charged at the full output rate (no cache discount on generated text).
Tiered rows (context-size / peak / off-peak) are listed separately as in the official table. Beam Free = $0 (limited time).
Prices are built into this page, so it works offline — just open the file.
Press “↻ Refresh from source” to fetch the latest Go-plan table from the
OpenCode Go pricing docs;
the table updates in place and fresh prices are remembered for an hour.
To update the built-in snapshot itself, edit the MODELS list in index.html.
“Tokens / sub” is the total tokens the subscription’s allowance buys at the current hit rate and output share (allowance ÷ blended $/1M price; updates when you change either).
Form selections (search, tokens, hit rate, output %, free-models toggle, column visibility, sort column and direction) are remembered in this browser via localStorage.