Token prices tracked across major providers, updated every 6 hours from OpenRouter and LiteLLM. The rates Tally uses when computing cost estimates.
Tally fetches prices every 6 hours from two sources: OpenRouter's /models API and LiteLLM's community-maintained price file. The displayed rate is the blended average.
Every routing decision and telemetry event uses the live rates to compute a cost estimate. The bandit's reward function weighs cost against quality — cheaper models win when quality holds up.
When both sources return a price, Tally compares them. A green check means they agree within 10%. An amber warning means a significant discrepancy — worth double-checking before budgeting.
Tally handles the selection. You handle the work.
Next up
MCP Market →Public MCP servers tracked in the Tally ecosystem — usage data, reliability, and more.