Sixteen interactive tools built on the same rate cards Continuum uses to price real usage: current per-token API rates, real plan prices, and the cap arithmetic the vendors publish as multipliers instead of numbers. Everything runs in your browser. Nothing is uploaded.
Compare a workload across 35 verified models: input, output, cache share, requests per day. Monthly bill per model, side by side.
Rates checked August 2026 → counterPaste text, get a token estimate, the cost at any current model, and whether it fits the context window. Nothing leaves your browser.
Estimate + per-model cost →Pro, Max 5x, Max 20x, Fable 5 in the mix, stacked accounts, or the API: the plan that survives your week, with the cap math shown.
Find the Claude plan → pickerPlus, Pro, Business seats, flex credits, or the API: which OpenAI plan survives how you actually run Codex.
Find the Codex plan → calculatorSubscription plus uncapped grok-4.6 overage, with the undisclosed weekly allowance shown as a low-to-high range.
Estimate the Grok Bot bill → pickerSix plans include Grok Bot and two lookalikes do not. Four questions, your cheapest route in, and the link you cannot undo.
Find the Grok Bot plan → identifyGrok, Grok Bot, Grok Automations, Grok Build, or the community grok-cli: three questions, the price of each, and which ones have an API.
Tell the five apart →Seats times plans times real usage patterns: the bill, the cost per developer, and where the idle-seat waste hides.
Budget the team → budgetBuild next quarter’s AI forecast by team, watch the threshold crossings, export the budget memo.
Plan the quarter → capsSet a per-seat cap and see exactly what gets blocked, for whom, on which day. Saved versus blocked, quantified.
Simulate the cap → limitsHow many concurrent agents your plans and API tiers actually support before the 429, and which meter fires first.
Plan the headroom → auditPaste a Copilot or generic seat export, get the reclaim list and the dollars back. Runs entirely in your browser.
Audit the seats → financeTurn usage rows into a finance-ready allocation: chargeback or showback, with CSV and Markdown exports.
Allocate the spend → tcoSelf-host on GPUs or buy tokens: the break-even curve for your volume, with every assumption stated.
Find the break-even → policyRoles times providers times data rules, conflict checks included. Exports policy JSON and a Markdown doc.
Build the policy → reviewWeighted, provenance-labeled scoring for model approvals, fed by independent benchmarks. Exports the decision memo.
Score the candidates →AI pricing is deliberately hard to compare: per-token rates on one page, plan multipliers on another, cache discounts in a footnote. These tools carry one curated rate table, verified against every provider's live pricing page and stamped with the date we checked. When a number is an estimate, the page says so and shows the assumption.
They are also the static version of what Continuum does live: the desktop app reads the agent history already on your machine and prices every real token by repo, model, and day. When the calculator's guess gets interesting, the scan replaces it with your actual number in about a minute.