GPT-Live 1 Pricing (2026): Real Cost Per Minute

Session time

$0.05/min

billed per second

On our stack

$54.55

$0.0546 at 1,000 min, excl. tokens

Concurrency

Build tier sessions

Telephony

Bring your own

GPT-Live 1 pricing at a glance

GPT-Live 1 pricing summary, USD

GPT-Live 1 pricing summary, USD, checked Oct 29, 2026
Item	Value
Session time	$0.05/min
Billing increment	per second, session duration, no rounding up to a whole minute
Audio tokens	No token charge
Backend model and tool calls	Backend model (Responses delegation) and tool usage billed at normal model/tool rates
Suggested backend model	gpt-6-luna (or gpt-6-sol for complex tasks)
gpt-6-luna tokens	$0.10 in, $0.01 cached, $0.50 out per 1M tokens
Free tier	None
Concurrency	Build 50, Launch 300, Grow 500 concurrent sessions; Free not supported
Generally available	Sep 10, 2026
Telephony	Bring your own: Twilio Agent Connect, Telnyx, Bandwidth (SIP), LiveKit, Daily/Pipecat
All-in at 1,000 min (Twilio SIP, one number, excl. backend tokens)	$54.55 $0.0546/min

Sources: developers.openai.com · Oct 29, 2026; developers.openai.com · Oct 29, 2026; developers.openai.com · Oct 29, 2026; developers.openai.com · Oct 29, 2026; developers.openai.com · Oct 29, 2026

What a minute on GPT-Live 1 actually costs

Cost per minute = GPT-Live 1 session time + your carrier per-minute rate + backend model tokens + tool calls. The carrier phone-number fee is monthly. Audio has no token charge. The table below leaves backend tokens out, so the real cost is higher.

Session time

GPT-Live 1 went generally available on Sep 10, 2026. Voice sessions are $0.05/min, and per second, session duration, no rounding up to a whole minute.

Backend model

Backend model (Responses delegation) and tool usage billed at normal model/tool rates. Delegation suggests gpt-6-luna (or gpt-6-sol for complex tasks). gpt-6-luna list price is $0.10 in, $0.01 cached, $0.50 out per 1M tokens. Per-minute token use is not published, so it is not in the totals.

Tool calls

Tool usage is billed at the normal tool rates. OpenAI does not publish a GPT-Live tool price on the session line, so tool calls stay out of the tables.

Telephony

Bring your own: Twilio Agent Connect, Telnyx, Bandwidth (SIP), LiveKit, Daily/Pipecat. OpenAI publishes no per-minute telephony charge for GPT-Live. Twilio Elastic SIP charges $0.0034/min for US local inbound and $1.15/month for a local number.

GPT-Live 1 plans

The session rate is one price. Build, Launch and Grow change the concurrency cap, not the session rate. Free tier: None.

GPT-Live 1 concurrency limit

Build 50, Launch 300, Grow 500 concurrent sessions; Free not supported. A 10-call setup fits on Build, which includes 50 concurrent sessions. No per-session fee is published.

GPT-Live 1 cost at 500 to 10,000 minutes a month

Scenario (our assumptions, not vendor figures): inbound calls on one US local number, average call 3 minutes, and the business needs 10 calls at once. The carrier is a Twilio Elastic SIP trunk. Backend model tokens and tool calls are excluded. The real cost is higher.

Monthly cost · USD

Our calculation. GPT-Live 1 monthly cost by volume, USD, checked Oct 29, 2026. Stack: session time, Twilio Elastic SIP Trunking, one US local number, 10 concurrent calls. Backend tokens and tool calls are excluded, so the real cost is higher.
Line item	500 min	1k min	5k min	10k min
GPT-Live 1 session time ($0.05/min)	$25.00	$50.00	$250.00	$500.00
Telephony: Twilio Elastic SIP, US local inbound	$1.70	$3.40	$17.00	$34.00
Twilio local number	$1.15	$1.15	$1.15	$1.15
Concurrency for 10 calls (Build includes 50)	$0.00	$0.00	$0.00	$0.00
Backend tokens (per-minute use not published)	Not published	Not published	Not published	Not published
Monthly total (excl. backend tokens)	$27.85	$54.55	$268.15	$535.15
Effective per minute	$0.0557	$0.0546	$0.0536	$0.0535

Sources: developers.openai.com · Oct 29, 2026; twilio.com · Oct 28, 2026; developers.openai.com · Oct 29, 2026

Math: minutes × ($0.05/min + $0.0034/min) + $1.15/month. At 1,000 minutes that is $54.55, or $0.0546 a minute. Run your own numbers.

Hidden costs: backend tokens, tools, carrier extras

Backend model tokens are billed separately. Per-minute use is not published, so the real cost is higher than the table.
gpt-6-luna is $0.10 in, $0.01 cached, $0.50 out per 1M tokens.
Tool calls are billed at normal tool rates. No GPT-Live tool price is in the session line.
Bring your own telephony. The Twilio lines above are the carrier, not an OpenAI fee.

GPT-Live 1 price changes

GPT-Live 1 price changes recorded on this site

GPT-Live 1 price changes recorded on this site
Date	Change
Sep 10, 2026	GPT-Live 1 is generally available in the API.
Oct 29, 2026	First recorded. Session time is $0.05/min, billed per second.

Sources: developers.openai.com · Oct 29, 2026; developers.openai.com · Oct 29, 2026

Is GPT-Live 1 worth it?

GPT-Live 1 suits teams that want a per-second session rate and can bring a SIP trunk. At $0.0546 a minute on our stack, the session-plus-carrier total is $54.55 for 1,000 minutes, and that excludes backend tokens, so the real cost is higher. The session rate is $0.05/min, against Mistral Voxtral Small 4 at $0.04/min and Grok Voice at $0.08/min. Compare more in Vapi alternatives.

FAQ

How much does GPT-Live 1 cost per minute?

$0.05/min of session time, billed per second. Audio has no token charge. On a Twilio Elastic SIP trunk with one number, 1,000 minutes are $54.55. That excludes backend tokens, so the real cost is higher. Source: https://developers.openai.com/api/docs/models/gpt-live-1 (checked Oct 29, 2026).

When did GPT-Live 1 launch?

GPT-Live 1 went generally available on Sep 10, 2026. Source: https://developers.openai.com/api/docs/changelog (checked Oct 29, 2026).

Is GPT-Live 1 free?

No. The rate-limit table lists the free tier as None. Source: https://developers.openai.com/api/docs/models/gpt-live-1 (checked Oct 29, 2026).

Do I pay for the model separately?

Yes. Backend model and tool calls are billed separately. Delegation suggests gpt-6-luna (or gpt-6-sol for complex tasks). gpt-6-luna is $0.10 in, $0.01 cached, $0.50 out per 1M tokens. Per-minute token use is not published. Source: https://developers.openai.com/api/docs/pricing (checked Oct 29, 2026).

What is the GPT-Live 1 concurrency limit?

Build 50, Launch 300, Grow 500 concurrent sessions; Free not supported. Source: https://developers.openai.com/api/docs/models/gpt-live-1 (checked Oct 29, 2026).

Does OpenAI include a phone number?

No. Bring your own: Twilio Agent Connect, Telnyx, Bandwidth (SIP), LiveKit, Daily/Pipecat. Source: https://developers.openai.com/api/docs/guides/live-partner-integrations (checked Oct 29, 2026).