Independent planning tool · updated July 24, 2026

GPT Live Pricing Calculator

Use this GPT Live pricing calculator to estimate the cost of real-time voice conversations. Enter audio minutes for a planning estimate or use actual input, cached input, output, and text token counts for an exact calculation based on OpenAI's published GPT-Realtime-2.1 rates. You can also compare ChatGPT Free, Go, Plus, Pro $100, and Pro $200 Live limits over a rolling 24-hour period. ChatGPT subscriptions and API usage are calculated separately because they are independent products.

Compare official Live limits

Live usage is measured over a rolling 24-hour period. Choose the intelligence mode you expect to use.

Official limits, verified July 24, 2026: Limits are rolling, can change, and ChatGPT notifies you when reached. A single Live conversation can last up to two hours. Free's mini limit is not published as a fixed number. Go pricing may vary by country. Source: OpenAI ChatGPT Voice.

Method and assumptions

How the GPT Live Pricing Calculator Works

The page contains two calculators for two different buying decisions. The Consumer Plans tab estimates how much ChatGPT Live time you use in a rolling 24-hour window and compares that figure with published plan limits. The Realtime API tab estimates developer API spend. A ChatGPT subscription does not include Realtime API credits, and buying a higher ChatGPT plan does not reduce a separate API invoice.

Estimate Cost from Audio Minutes

Minute mode is useful before an application has real usage data. It converts each audio input minute to an estimated 600 audio tokens and each audio output minute to an estimated 1,200 audio tokens, then applies the selected model's per-million-token prices. OpenAI's current pricing page does not publish a fixed audio-token-per-minute conversion, so these factors are planning assumptions rather than official rates. Actual token volume can change with silence, turn length, truncation, and how much audio the model generates.

For planning, the GPT Live pricing calculator separates input and output because generated speech is priced differently from user audio. Enter the number of minutes your users speak under audio input and the number of minutes the model speaks under audio output. Do not enter total call duration in both fields unless both sides actually speak for that full duration.

Calculate Exact Cost from API Tokens

Token mode is the appropriate choice for billing reconciliation. Copy the actual uncached audio input, cached audio input, audio output, text input, cached text input, and text output token counts from your API usage records. The calculator multiplies each category by the corresponding official rate and adds the results. Keeping cached and uncached input separate matters because cached input is substantially cheaper.

The result covers model token charges represented by the fields on this page. It does not estimate unrelated infrastructure, telephony, storage, network, transcription, tool calls, or taxes. Use the OpenAI billing dashboard and your invoice as the final record of charges.

Official token rates

Current GPT Live API Pricing

OpenAI bills the Realtime API by tokens rather than by a universal minute price. The audio rates below are built into this calculator and were checked against the official OpenAI API pricing page on July 24, 2026. Prices are in US dollars per one million tokens. Recheck the source before committing a production budget because OpenAI can update models and rates.

Audio token rates used by the calculator
Model	Audio input	Cached input	Audio output
GPT-Realtime-2.1	$32.00 / 1M	$0.40 / 1M	$64.00 / 1M
GPT-Realtime-2.1 mini	$10.00 / 1M	$0.30 / 1M	$20.00 / 1M

GPT-Realtime-2.1 Pricing

For the full model, text tokens are $4.00 per million uncached input tokens, $0.40 per million cached input tokens, and $24.00 per million output tokens. Audio is $32.00 per million uncached input tokens, $0.40 per million cached input tokens, and $64.00 per million output tokens.

GPT-Realtime-2.1 Mini Pricing

For the mini model, text tokens are $0.60 per million uncached input tokens, $0.06 per million cached input tokens, and $2.40 per million output tokens. Audio is $10.00 per million uncached input tokens, $0.30 per million cached input tokens, and $20.00 per million output tokens.

Consumer subscriptions

Compare ChatGPT Live Plan Limits

The Consumer Plans tab estimates daily voice time by multiplying conversations per day by average minutes per conversation. It compares the result with the rolling limits published in OpenAI's ChatGPT Voice article. These are usage windows, not monthly minute bundles, and a single Live conversation can last up to two hours. OpenAI may change limits, apply guardrails, or show a different local Go price.

Free, Go, and Plus

Free includes a variable amount of GPT-Live-1 mini access; OpenAI does not publish a fixed number that this tool can safely promise. Go is shown at the $8 reference price, while local pricing can vary. Go and Plus currently publish one rolling hour for each GPT-Live-1 intelligence bucket and two rolling hours for GPT-Live-1 mini. Plus is $20 per month and also differs from Go in features beyond voice, which this calculator does not score.

Pro $100 and Pro $200

Pro at $100 per month currently publishes 12 rolling hours for each GPT-Live-1 intelligence bucket and 24 rolling hours for mini. Pro at $200 per month publishes unlimited GPT-Live-1 access subject to abuse guardrails. The recommendation highlights the lowest selected subscription whose published voice limit covers the estimate. It is not a claim that one plan is best overall.

Worked calculations

GPT Live Pricing Calculator Examples

These examples show the exact arithmetic used in minute and token modes. They make the assumptions visible so you can reproduce every result instead of relying on an unexplained total.

Example: 500 Input and 500 Output Minutes

At the planning conversion used here, 500 input minutes equal an estimated 300,000 input audio tokens, while 500 output minutes equal an estimated 600,000 output audio tokens. For GPT-Realtime-2.1, input costs 300,000 / 1,000,000 × $32 = $9.60. Output costs 600,000 / 1,000,000 × $64 = $38.40. The estimated total is $48.00.

With GPT-Realtime-2.1 mini, the same estimated token counts cost $3.00 for input and $12.00 for output, for an estimated total of $15.00. Text tokens and cached audio are zero in this example.

Example: Calculating Cost from Actual Tokens

Suppose your GPT-Realtime-2.1 usage record contains 1,000,000 uncached audio input tokens, 500,000 cached audio input tokens, and 750,000 audio output tokens. The audio calculation is $32.00 + $0.20 + $48.00, producing a total of $80.20 before any text token charges.

Enter those values in exact token mode, then add text usage in its matching categories. This method avoids the minute conversion and is the better way to audit recorded usage against billing. The GPT Live pricing calculator will update each cost bar as the token counts change.

Keep the products separate

ChatGPT Live Subscription vs Realtime API Pricing

ChatGPT Free, Go, Plus, and Pro are consumer subscriptions with product features and Live usage limits. GPT-Realtime-2.1 and GPT-Realtime-2.1 mini are developer API models billed from token usage. A $100 or $200 Pro subscription is not an API allowance, and the API does not become free after a ChatGPT plan reaches a supposed break-even point. There is therefore no valid “Plus overage” or “Pro versus API” break-even calculation to display.

Choose a ChatGPT plan based on the published Live limits and the rest of the subscription features. Estimate an application separately from expected API tokens, model choice, caching, and output length. Keeping those decisions separate is the most reliable way to avoid underestimating costs.

Answers before you buy

GPT Live Pricing Calculator FAQ

How much does GPT Live cost?

ChatGPT Live is included with eligible subscriptions. The separate GPT-Realtime-2.1 API costs $32 per million uncached audio input tokens, $0.40 per million cached audio input tokens, and $64 per million audio output tokens. GPT-Realtime-2.1 mini costs $10, $0.30, and $20 per million tokens for the same categories.

What is the difference between GPT Live Free and Plus?

Free receives a variable amount of GPT-Live-1 mini access, so no fixed minute allowance can be guaranteed. Plus costs $20 per month and currently publishes one rolling hour for each GPT-Live-1 intelligence bucket plus two rolling hours of mini. Limits can change and are separate from API billing.

How is GPT Live API pricing calculated?

For exact billing, multiply the actual uncached input, cached input, and output token counts in your API usage by the official per-million-token rates for the selected model. Minute mode is only a planning estimate because audio token usage can vary.

Is GPT Live Pro worth it?

The $100 Pro tier currently publishes 12 rolling hours for each GPT-Live-1 intelligence bucket and 24 rolling hours for mini. The $200 tier offers unlimited GPT-Live-1 subject to guardrails. Whether it is worth it depends on Live usage and other ChatGPT features, not API savings, because API usage is billed separately.

Explore our guides

How to enableSet up voice on web, iOS, and Android.FeaturesSee what changes in the voice experience.GPT-Live APIDeveloper context and API coverage.Pricing guidePlan-by-plan editorial breakdown.LanguagesLanguage support and voice quality.Free tierWhat you can expect for $0.