Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
70 changes: 70 additions & 0 deletions src/server/providers/fixtures/openai-gpt-live-1.1.md.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,70 @@
# GPT-Live-1.1

> For the complete documentation index, see [llms.txt](/llms.txt). Markdown versions of documentation pages are available by appending `.md` to the page URL.

> Full-duplex voice model for multilingual conversations.

Model ID: `gpt-live-1.1`

GPT-Live-1.1 can listen while it speaks. Compared with GPT-Live-1, it transcribes speech more accurately and maintains a more consistent accent and voice across languages. Describe in plain language when the model should hand a task to your backend for reasoning or tool use.

Learn how to [start a voice session](/api/docs/guides/live), [prompt GPT-Live](/api/docs/guides/live-prompting), and [configure delegation](/api/docs/guides/live-delegation).

## Model details

- Default snapshot: `gpt-live-1.1`
- Input modalities: audio, text
- Output modalities: audio, text

## Pricing

Voice sessions are billed by duration. Backend model and tool usage is billed separately.

### Live session duration

| Metric | Price | Unit |
| --- | ---: | --- |
| Active | $0.05 | minute |
| Standby | $0.02 | minute |

- The lower standby rate applies to WebRTC sessions while inference is paused. See [GPT-Live pricing](/api/docs/pricing#live-session-duration).

## Endpoints

| Endpoint | Route | Support |
| --- | --- | --- |
| Live | `v1/live/sessions` | Supported |
| Chat Completions | `v1/chat/completions` | Not supported |
| Responses | `v1/responses` | Not supported |
| Realtime | `v1/realtime` | Not supported |
| Realtime translation | `v1/realtime/translations` | Not supported |
| Realtime transcription | `v1/realtime/transcription_sessions` | Not supported |
| Assistants | `v1/assistants` | Not supported |
| Batch | `v1/batch` | Not supported |
| Fine-tuning | `v1/fine-tuning` | Not supported |
| Embeddings | `v1/embeddings` | Not supported |
| Image generation | `v1/images/generations` | Not supported |
| Videos | `v1/videos` | Not supported |
| Image edit | `v1/images/edits` | Not supported |
| Speech generation | `v1/audio/speech` | Not supported |
| Transcription | `v1/audio/transcriptions` | Not supported |
| Translation | `v1/audio/translations` | Not supported |
| Moderation | `v1/moderations` | Not supported |
| Completions (legacy) | `v1/completions` | Not supported |

## Supported features

- streaming
- function_calling

## Snapshots

Use `gpt-live-1.1` in your API requests.

- `gpt-live-1.1`

## Rate limits

GPT-Live-1.1 and GPT-Live-1 have separate concurrent-session limits. Each model’s limit applies across inference regions.

See your [organization’s limits](https://platform.openai.com/settings/organization/limits) for the rate limits that apply to your account.
88 changes: 88 additions & 0 deletions src/server/providers/fixtures/openai-gpt-live-1.md.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,88 @@
# GPT-Live-1

> For the complete documentation index, see [llms.txt](/llms.txt). Markdown versions of documentation pages are available by appending `.md` to the page URL.

> Our premier model for natural, expressive voice conversations with smooth interruption handling.

Model ID: `gpt-live-1`

GPT-Live-1 is a full-duplex voice model for real-time conversations. It can listen and speak at the same time, and delegate reasoning and tool use to a backend agent.

Learn how to [get started with GPT-Live](/api/docs/guides/live), [configure delegation](/api/docs/guides/live-delegation), and [prompt the model](/api/docs/guides/live-prompting).

## Model details

- Default snapshot: `gpt-live-1`
- Input modalities: audio, text
- Output modalities: audio, text
- Unsupported modalities: image, video
- Jul 31, 2025 knowledge cutoff

## Pricing

Voice sessions are billed per second. Backend model and tool usage is billed separately.

### Live session duration

| Metric | Price | Unit |
| --- | ---: | --- |
| Active | $0.05 | minute |
| Standby | $0.02 | minute |

- The lower standby rate applies to WebRTC sessions while inference is paused. See [GPT-Live pricing](/api/docs/pricing#live-session-duration).
- Session duration is not rounded up to the next whole minute.
- Backend Responses calls use the normal pricing for the configured model and tools.

## Endpoints

| Endpoint | Route | Support |
| --- | --- | --- |
| Live | `v1/live/sessions` | Supported |
| Chat Completions | `v1/chat/completions` | Not supported |
| Responses | `v1/responses` | Not supported |
| Realtime | `v1/realtime` | Not supported |
| Realtime translation | `v1/realtime/translations` | Not supported |
| Realtime transcription | `v1/realtime/transcription_sessions` | Not supported |
| Assistants | `v1/assistants` | Not supported |
| Batch | `v1/batch` | Not supported |
| Fine-tuning | `v1/fine-tuning` | Not supported |
| Embeddings | `v1/embeddings` | Not supported |
| Image generation | `v1/images/generations` | Not supported |
| Videos | `v1/videos` | Not supported |
| Image edit | `v1/images/edits` | Not supported |
| Speech generation | `v1/audio/speech` | Not supported |
| Transcription | `v1/audio/transcriptions` | Not supported |
| Translation | `v1/audio/translations` | Not supported |
| Moderation | `v1/moderations` | Not supported |
| Completions (legacy) | `v1/completions` | Not supported |

## Supported features

- streaming
- function_calling

## Unsupported features

- structured_outputs
- fine_tuning
- predicted_outputs

## Snapshots

Use `gpt-live-1` in your API requests.

- `gpt-live-1`

## Rate limits

Rate limits are measured in concurrent sessions.

Unsupported usage tiers: Free.

### Concurrent sessions

| Tier | Concurrent sessions |
| --- | ---: |
| Build | 50 |
| Launch | 300 |
| Grow | 500 |
80 changes: 79 additions & 1 deletion src/server/providers/model-facts.test.ts
Original file line number Diff line number Diff line change
@@ -1,3 +1,4 @@
import { readFileSync } from 'node:fs'
import { describe, expect, it, vi } from 'vitest'

import { anthropicCapabilities } from './anthropic.ts'
Expand Down Expand Up @@ -35,9 +36,10 @@ import {
} from './model-facts.ts'
import { price } from '@modelschemas/rate-card'
import { evaluateCard } from '@modelschemas/rate-card/evaluator'
import { parseStoredRateCard } from '../rate-card.ts'
import { parseStoredRateCard, projectTokenPricing } from '../rate-card.ts'
import { openAiCompatModelFacts } from './openai-compat.ts'
import {
compileLiveSessionCard,
pageSlugFor,
parseModelIndex,
parseModelPage,
Expand Down Expand Up @@ -855,6 +857,82 @@ Landscape: 1920x1080 | $0.7 | second |
})
})

it('prices active and standby seconds from both gpt-live pages', () => {
const source = {
url: 'https://developers.openai.com/api/docs/models/gpt-live-1.md',
hash: 'a'.repeat(64),
extractedAt: '2026-11-20T00:00:00.000Z',
}
for (const file of [
'openai-gpt-live-1.md.txt',
'openai-gpt-live-1.1.md.txt',
]) {
const page = parseModelPage(
readFileSync(new URL(`./fixtures/${file}`, import.meta.url), 'utf8'),
)
expect(page?.pricing?.session).toEqual({
activeSeconds: 0.05 / 60,
standbySeconds: 0.02 / 60,
})
expect(page?.pricing?.unit).toBeUndefined()
const card = page?.pricing?.session
? compileLiveSessionCard(page.pricing.session, source)
: null
if (!card) throw new Error(`${file}: did not compile`)
expect(parseStoredRateCard(card)).toEqual(card)
expect(projectTokenPricing(card)).toEqual({
currency: 'USD',
per: 'second',
})
// A minute is not rounded up: 30s is half the per-minute rate.
expect(
price(card, {}, { active_seconds: 60, standby_seconds: 0 }),
).toBeCloseTo(0.05, 9)
expect(
price(card, {}, { active_seconds: 0, standby_seconds: 60 }),
).toBeCloseTo(0.02, 9)
expect(
price(card, {}, { active_seconds: 30, standby_seconds: 30 }),
).toBeCloseTo(0.035, 9)
expect(() => price(card, {}, { active_seconds: 60 })).toThrow(
expect.objectContaining({ code: 'bad-input' }),
)
}
})

it('refuses a live table that drops either rate or names another metric', () => {
const live = (rows: string) =>
parseModelPricing(
pricing(`### Live session duration

| Metric | Price | Unit |
| --- | ---: | --- |
${rows}
`),
)
expect(
live(`| Active | $0.05 | minute |
| Standby | $0.02 | minute |`),
).toMatchObject({
session: { activeSeconds: 0.05 / 60, standbySeconds: 0.02 / 60 },
})
expect(live('| Active | $0.05 | minute |')).toBeNull()
expect(live('| Standby | $0.02 | minute |')).toBeNull()
expect(
live(`| Active | $0.05 | minute |
| Listening | $0.02 | minute |`),
).toBeNull()
expect(
live(`| Active | $0.05 | minute |
| Standby | $0.02 | minute |
| Extra | $0.01 | minute |`),
).toBeNull()
expect(
live(`| First | $0.05 | minute |
| Second | $0.02 | minute |`),
).toBeNull()
})

it('refuses a page that mixes a token table with a per-unit one', () => {
expect(
parseModelPricing(
Expand Down
Loading
Loading