Gemini 4 agent-run cost simulator over MCP: true per-turn totals vs naive estimates.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Turn-by-turn cost of a Gemini 4 agent run: every turn re-sends the growing context as input, thinking tokens are billed at the output rate. Returns the rates in force on the chosen date, the simulated total, the naive first-turn × turns estimate and their multiplier. Before the 2026-12-01 price change it also returns afterIncrease: the same run priced at the announced 2026-12-01 rates.
Number of turns (separate API calls) in the run.
Gemini 4 model to price.
Thinking budget. Thinking tokens are billed at the output price: off = 0×, low ≈0.5×, medium ≈1.5×, high ≈3× the visible output tokens.
Visible answer tokens per turn, excluding thinking.
Net history added after each completed turn; billed from turn 2 onward.
System prompt, tool definitions and documents sent on turn 1.
New user/tool content billed on top of the context each turn.
ISO date (YYYY-MM-DD) that selects the price segment in force that day. Defaults to today (UTC).
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http geminicost https://geminicost.craftstead.app/mcp
Turn-by-turn cost of a Gemini 4 agent run: every turn re-sends the growing context as input, thinking tokens are billed at the output rate. Returns the rates in force on the chosen date, the simulated total, the naive first-turn × turns estimate and their multiplier. Before the 2026-12-01 price change it also returns afterIncrease: the same run priced at the announced 2026-12-01 rates.
| Argument | Type | Description |
|---|---|---|
| date | string | ISO date (YYYY-MM-DD) that selects the price segment in force that day. Defaults to today (UTC). |
| turns* | number | Number of turns (separate API calls) in the run. |
| modelId* | string | Gemini 4 model to price. |
| thinkingLevel* | string | Thinking budget. Thinking tokens are billed at the output price: off = 0×, low ≈0.5×, medium ≈1.5×, high ≈3× the visible output tokens. |
| outputTokensPerTurn* | number | Visible answer tokens per turn, excluding thinking. |
| contextGrowthPerTurn* | number | Net history added after each completed turn; billed from turn 2 onward. |
| initialContextTokens* | number | System prompt, tool definitions and documents sent on turn 1. |
| userToolInputTokensPerTurn* | number | New user/tool content billed on top of the context each turn. |
{
"type": "object",
"required": [
"modelId",
"thinkingLevel",
"turns",
"initialContextTokens",
"contextGrowthPerTurn",
"userToolInputTokensPerTurn",
"outputTokensPerTurn"
],
"properties": {
"date": {
"type": "string",
"pattern": "^\\d{4}-\\d{2}-\\d{2}$",
"examples": [
"2026-11-20",
"2026-12-05"
],
"description": "ISO date (YYYY-MM-DD) that selects the price segment in force that day. Defaults to today (UTC)."
},
"turns": {
"type": "number",
"maximum": 365,
"minimum": 1,
"description": "Number of turns (separate API calls) in the run."
},
"modelId": {
"enum": [
"gemini-4-hydrogen",
"gemini-4-carbon",
"gemini-4-argon"
],
"type": "string",
"description": "Gemini 4 model to price."
},
"thinkingLevel": {
"enum": [
"off",
"low",
"medium",
"high"
],
"type": "string",
"description": "Thinking budget. Thinking tokens are billed at the output price: off = 0×, low ≈0.5×, medium ≈1.5×, high ≈3× the visible output tokens."
},
"outputTokensPerTurn": {
"type": "number",
"minimum": 0,
"description": "Visible answer tokens per turn, excluding thinking."
},
"contextGrowthPerTurn": {
"type": "number",
"minimum": 0,
"description": "Net history added after each completed turn; billed from turn 2 onward."
},
"initialContextTokens": {
"type": "number",
"minimum": 0,
"description": "System prompt, tool definitions and documents sent on turn 1."
},
"userToolInputTokensPerTurn": {
"type": "number",
"minimum": 0,
"description": "New user/tool content billed on top of the context each turn."
}
},
"additionalProperties": false
}Price segment in force on a date for every recorded Gemini 4 model, or for one model: input and output USD per 1,000,000 tokens, the segment start and end dates and its status. Segments marked announced-unpriced have no published values and are reported as such.
| Argument | Type | Description |
|---|---|---|
| date | string | ISO date (YYYY-MM-DD) that selects the price segment in force that day. Defaults to today (UTC). |
| modelId | string | Restrict the answer to one model. Omit for all recorded models. |
{
"type": "object",
"properties": {
"date": {
"type": "string",
"pattern": "^\\d{4}-\\d{2}-\\d{2}$",
"examples": [
"2026-11-20",
"2026-12-05"
],
"description": "ISO date (YYYY-MM-DD) that selects the price segment in force that day. Defaults to today (UTC)."
},
"modelId": {
"enum": [
"gemini-4-hydrogen",
"gemini-4-carbon",
"gemini-4-argon"
],
"type": "string",
"description": "Restrict the answer to one model. Omit for all recorded models."
}
},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 18, 2026.
No readOnlyHint / destructiveHint annotations, so clients can't tell a read from a write without asking.
Fix: Add annotations to each tool, e.g. { readOnlyHint: true } for tools that only read.