docs: update OpenAI model examples from gpt-4o to gpt-5.2 #222
Conversation
📝 WalkthroughWalkthroughDocumentation-only updates: replaced model identifier strings from gpt-4o/gpt-4o-mini to gpt-5.2/gpt-5-mini across examples in guides, API, protocol, and reference docs. No functional code, signatures, or control flow were modified. (≈26 files) Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 inconclusive)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Update documentation examples to use gpt-5.2 instead of gpt-4o, and gpt-5-mini instead of gpt-4o-mini for summarization examples. Audio/transcription docs were intentionally left unchanged as they reference specialized audio models (gpt-4o-audio-preview, whisper-1, etc.) that have different naming conventions.
shkumbinhasani
force-pushed
the
docs/update-openai-model-examples
branch
from
6272491 to
e6e670e
Compare
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In @docs/guides/tree-shaking.md:
- Line 46: The docs reference an invalid OpenAI model identifier `gpt-5.2-mini`;
update each occurrence (e.g., the adapter call like openaiText('gpt-5.2-mini')
at lines noted) to a valid GPT-5.2 variant such as `gpt-5.2-instant` (or
`gpt-5.2-thinking` / `gpt-5.2-pro`) so all adapter entries and examples
consistently use a supported model name.
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (30)
docs/adapters/openai.mddocs/api/ai.mddocs/getting-started/overview.mddocs/getting-started/quick-start.mddocs/guides/agentic-cycle.mddocs/guides/client-tools.mddocs/guides/migration.mddocs/guides/multimodal-content.mddocs/guides/runtime-adapter-switching.mddocs/guides/server-tools.mddocs/guides/streaming.mddocs/guides/structured-outputs.mddocs/guides/tool-approval.mddocs/guides/tool-architecture.mddocs/guides/tools.mddocs/guides/tree-shaking.mddocs/protocol/chunk-definitions.mddocs/protocol/http-stream-protocol.mddocs/protocol/sse-protocol.mddocs/reference/functions/chat.mddocs/reference/functions/combineStrategies.mddocs/reference/functions/maxIterations.mddocs/reference/functions/streamToText.mddocs/reference/functions/summarize.mddocs/reference/functions/toHttpResponse.mddocs/reference/functions/toHttpStream.mddocs/reference/functions/toServerSentEventsResponse.mddocs/reference/functions/untilFinishReason.mddocs/reference/interfaces/SummarizeAdapter.mddocs/reference/interfaces/TextAdapter.md
🧰 Additional context used
🧠 Learnings (9)
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Implement framework integrations using the headless `tanstack/ai-client` for state management with framework-specific hooks (useChat) on top
Applied to files:
docs/getting-started/overview.mddocs/reference/functions/chat.mddocs/api/ai.mddocs/guides/tools.mddocs/adapters/openai.mddocs/guides/multimodal-content.mddocs/guides/migration.mddocs/guides/tree-shaking.mddocs/guides/streaming.mddocs/guides/client-tools.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to **/*.{ts,tsx} : Implement stream processing with StreamProcessor for handling chunked responses and support partial JSON parsing for streaming AI responses
Applied to files:
docs/reference/functions/chat.mddocs/reference/functions/toHttpResponse.mddocs/api/ai.mddocs/adapters/openai.mddocs/protocol/chunk-definitions.mddocs/reference/functions/toServerSentEventsResponse.mddocs/reference/functions/untilFinishReason.mddocs/getting-started/quick-start.mddocs/protocol/sse-protocol.mddocs/guides/migration.mddocs/guides/server-tools.mddocs/reference/functions/toHttpStream.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to packages/typescript/*/src/adapters/*.ts : Create individual adapter implementations for each provider capability (text, embed, summarize, image) with separate exports to enable tree-shaking
Applied to files:
docs/guides/tools.mddocs/reference/functions/summarize.mddocs/guides/multimodal-content.mddocs/guides/migration.mddocs/guides/runtime-adapter-switching.mddocs/guides/tree-shaking.mddocs/reference/interfaces/TextAdapter.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to **/*.{ts,tsx} : Use type-safe per-model configuration with provider options typed based on selected model to ensure compile-time safety
Applied to files:
docs/reference/interfaces/SummarizeAdapter.mddocs/guides/tree-shaking.md
📚 Learning: 2025-12-27T21:39:29.563Z
Learnt from: harry-whorlow
Repo: TanStack/ai PR: 117
File: packages/typescript/ai-ollama/src/meta/model-meta-llama-guard3.ts:70-75
Timestamp: 2025-12-27T21:39:29.563Z
Learning: The standard template comments in ai-ollama model-meta files (like "Models with text, image, audio, video (no document)") should not be modified to match individual model capabilities, as they are intentionally kept uniform across all model-meta-*.ts files for consistency, regardless of what each specific model actually supports.
Applied to files:
docs/guides/multimodal-content.mddocs/reference/interfaces/TextAdapter.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Maintain type safety through multimodal content support (image, audio, video, document) with model capability awareness
Applied to files:
docs/guides/multimodal-content.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to **/*.{ts,tsx} : Use tree-shakeable adapter architecture for provider implementations - export specialized adapters (text, embedding, summarize, image) as separate imports from `/adapters` subpath rather than monolithic adapters
Applied to files:
docs/guides/migration.mddocs/guides/runtime-adapter-switching.mddocs/guides/tree-shaking.mddocs/reference/interfaces/TextAdapter.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to packages/typescript/*/src/index.ts : Export tree-shakeable adapters with clear subpath exports in package.json (e.g., `tanstack/ai/adapters`, `tanstack/ai-openai/adapters`) to minimize bundle size
Applied to files:
docs/guides/tree-shaking.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to **/*.{ts,tsx} : Implement isomorphic tool system using `toolDefinition()` with `.server()` and `.client()` implementations for dual-environment execution
Applied to files:
docs/guides/client-tools.md
🪛 markdownlint-cli2 (0.18.1)
docs/protocol/sse-protocol.md
82-82: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
88-88: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
🔇 Additional comments (25)
docs/reference/functions/maxIterations.md (1)
35-35: LGTM! Model update is correct.The update to
gpt-5.2is accurate and aligns with OpenAI's latest model offerings. The example correctly demonstrates themaxIterationsagent loop strategy with the current flagship model.docs/reference/functions/toHttpResponse.md (1)
45-45: LGTM! Model reference updated correctly.The update to
gpt-5.2is valid. The example accurately demonstrates converting a chat stream to an HTTP response format using the current OpenAI flagship model.docs/reference/functions/untilFinishReason.md (1)
35-35: LGTM! Documentation update is accurate.The model reference has been correctly updated to
gpt-5.2. The example properly illustrates theuntilFinishReasonstrategy with OpenAI's current flagship model.docs/guides/tools.md (1)
191-191: GPT-5.2 model is confirmed to exist and is publicly available.OpenAI released GPT-5.2 on December 11, 2025, and it is available via the OpenAI API for developers. The code examples referencing this model are accurate.
docs/reference/functions/toHttpStream.md (1)
45-45: The gpt-5.2 model reference is accurate. OpenAI released GPT-5.2 in December 2025 with the official model identifiersgpt-5.2,gpt-5.2-chat-latest, andgpt-5.2-pro. The code example in the documentation is valid and will work for users.docs/guides/tool-approval.md (1)
68-68: No action needed—GPT-5.2 is a valid model.OpenAI released GPT-5.2 on December 11, 2025, and it is available to developers via the API. The documentation reference is accurate.
docs/guides/agentic-cycle.md (1)
126-126: No issues found. The gpt-5.2 model was announced by OpenAI on December 11, 2025, and is available in the API. The documentation example is accurate.docs/reference/functions/streamToText.md (1)
38-38: LGTM! Model reference updated correctly.The example now uses the latest GPT-5.2 model, which is appropriate for documentation updates.
docs/reference/functions/summarize.md (1)
45-45: LGTM! Summarization examples updated correctly.All examples now consistently use
gpt-5-mini, which is the correct model name for the faster, more cost-efficient version of GPT-5. The updates are appropriate for summarization tasks.Also applies to: 54-54, 63-63, 71-71
docs/reference/interfaces/SummarizeAdapter.md (1)
16-16: LGTM! Type parameter example updated correctly.The TModel generic parameter example now references
gpt-5.2, keeping the documentation current with the latest OpenAI models.docs/adapters/openai.md (1)
22-22: Thegpt-5.2model identifier is valid and publicly available as of January 2026. OpenAI's API documentation confirms this model, along with variants likegpt-5.2-chat-latestandgpt-5.2-pro. The documentation examples are accurate.docs/guides/server-tools.md (1)
152-152: Model identifier update looks consistent.The changes from
gpt-4otogpt-5.2are applied consistently in both examples. Pending verification that thegpt-5.2model exists and is publicly available (see verification request in combineStrategies.md).Also applies to: 210-210
docs/guides/streaming.md (1)
18-18: Model identifier update looks consistent.The changes from
gpt-4otogpt-5.2are applied consistently in both streaming examples. Pending verification that thegpt-5.2model exists and is publicly available (see verification request in combineStrategies.md).Also applies to: 40-40
docs/reference/functions/combineStrategies.md (1)
36-36: No action needed — GPT-5.2 is confirmed as a publicly available OpenAI model (released December 11, 2025) and the documentation update is valid.docs/protocol/http-stream-protocol.md (2)
86-88: Documentation examples consistently updated.All NDJSON example chunks have been systematically updated to reference
gpt-5.2, maintaining consistency across different chunk types (content, tool_call, done, error).Also applies to: 94-95, 101-101, 133-137, 153-153, 363-365
184-184: Code examples updated to usegpt-5.2.The implementation examples correctly demonstrate the new model identifier in the
openaiText()adapter calls.Also applies to: 239-239
docs/api/ai.md (2)
55-55: Clarify model choice for summarization examples.The PR description states that "summarization examples should use
gpt-5-miniinstead ofgpt-4o-mini", but these lines usegpt-5.2instead ofgpt-5.2-mini. Please confirm whether:
- Summarization should use
gpt-5.2(the main model)- Summarization should use
gpt-5.2-mini(the smaller, cost-effective variant)- Or if the PR description should be updated to reflect the actual changes
Also applies to: 329-329
24-24: API documentation examples consistently updated.All code examples across different API functions (
chat(),toolDefinition(), SSE utilities, and usage examples) have been systematically updated to usegpt-5.2.Also applies to: 34-34, 102-102, 115-115, 143-143, 170-170, 194-194, 277-277, 283-283, 291-291, 313-313
docs/protocol/chunk-definitions.md (1)
19-19: The documentation is accurate.gpt-5.2is a valid, released OpenAI model (released December 2025) and represents the current frontier release as of January 2026. The model identifier follows OpenAI's standard naming convention, with multiple variants available (gpt-5.2-instant, gpt-5.2-thinking, gpt-5.2-pro, gpt-5.2-codex). No API compatibility changes are required for this model reference.docs/guides/client-tools.md (1)
105-105: LGTM - Model references updated consistently.The adapter model references have been updated consistently from
gpt-4otogpt-5.2in both the POST handler example and the hybrid tools examples. The changes align with the broader documentation updates across the PR.Also applies to: 300-303
docs/guides/multimodal-content.md (2)
61-61: LGTM - Adapter usage updated consistently.The adapter model references and accompanying comments have been updated consistently to reference
gpt-5.2throughout the multimodal content examples.Also applies to: 285-287
106-107: Remove references to non-existent model variants.The documentation claims support for
gpt-5.2-miniandgpt-5.2-audio-preview, but OpenAI's current model lineup does not include these variants. As of December 2025, the gpt-5.2 family consists of text + image-capable models (Instant, Thinking, Pro variants), not separate "mini" or "audio-preview" versions. Audio functionality is provided through distinct models likegpt-realtimeandgpt-4o-audio-preview, not throughgpt-5.2-audio-preview. Update the documentation to reflect the actual available models and their actual capabilities.Likely an incorrect or invalid review comment.
docs/protocol/sse-protocol.md (1)
77-77: No action needed—gpt-5.2 is a valid and available model.OpenAI released GPT-5.2 on December 11, 2025, and it is currently available via the API in multiple variants (Instant, Thinking, and Pro). The model references in the SSE protocol documentation are accurate.
docs/guides/runtime-adapter-switching.md (1)
25-25: Model references updated consistently.The model references have been updated from
gpt-4otogpt-5.2and fromgpt-4o-minitogpt-5.2-miniconsistently across all adapter examples. Pending verification of model availability (see previous comment).Also applies to: 43-47, 75-75, 135-135, 166-166, 182-182
docs/guides/migration.md (1)
33-33: Migration examples updated consistently.All migration examples have been updated to use
gpt-5.2instead ofgpt-4o. The before/after patterns clearly demonstrate the API changes while using the updated model identifiers. Pending verification of model availability.Also applies to: 45-45, 52-52, 126-126, 141-141, 160-160, 174-174, 200-200, 214-214, 231-238, 260-260, 280-280, 374-374, 401-401
|
|
||
| const stream = chat({ | ||
| adapter: openaiText('gpt-4o'), | ||
| adapter: openaiText('gpt-5.2'), |
There was a problem hiding this comment.
🧩 Analysis chain
🌐 Web query:
OpenAI GPT models available January 2026 gpt-5.2
💡 Result:
As of January 12, 2026 the latest OpenAI release is GPT‑5.2 (available as GPT‑5.2 Instant, GPT‑5.2 Thinking, and GPT‑5.2 Pro; API names include gpt-5.2-instant and gpt-5.2-thinking). Earlier GPT families (GPT‑5.1, GPT‑5) and GPT‑4.1 remain available to developers/users via the API/ChatGPT. [1][2][3]
Sources:
- OpenAI — “Introducing GPT‑5.2” (Dec 11, 2025). [1]
- OpenAI — GPT‑5 system card / GPT‑5.2 update (Dec 11, 2025). [2]
- Reuters coverage of GPT‑5.2 launch (Dec 11, 2025). [3]
Replace gpt-5.2-mini with a valid OpenAI model identifier.
The documentation correctly references gpt-5.2 at multiple lines, which is valid as of January 2026. However, gpt-5.2-mini (lines 151, 168) does not exist in OpenAI's official offerings. According to OpenAI's December 2025 announcement, the available GPT-5.2 variants are gpt-5.2-instant, gpt-5.2-thinking, and gpt-5.2-pro. Update these references to use one of the valid API model names.
🤖 Prompt for AI Agents
In @docs/guides/tree-shaking.md at line 46, The docs reference an invalid OpenAI
model identifier `gpt-5.2-mini`; update each occurrence (e.g., the adapter call
like openaiText('gpt-5.2-mini') at lines noted) to a valid GPT-5.2 variant such
as `gpt-5.2-instant` (or `gpt-5.2-thinking` / `gpt-5.2-pro`) so all adapter
entries and examples consistently use a supported model name.
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In @docs/guides/multimodal-content.md:
- Around line 106-107: The docs list a non-existent model
`gpt-5.2-audio-preview`; update the multimodal table so it only references valid
models: keep `gpt-5.2` and `gpt-5-mini` as text+image, and either remove the
`gpt-5.2-audio-preview` row entirely or replace it with a documented
audio-capable model such as `gpt-4o-audio-preview` (or another official
audio-preview model) and adjust the modalities column accordingly.
📜 Review details
Configuration used: defaults
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (30)
docs/adapters/openai.mddocs/api/ai.mddocs/getting-started/overview.mddocs/getting-started/quick-start.mddocs/guides/agentic-cycle.mddocs/guides/client-tools.mddocs/guides/migration.mddocs/guides/multimodal-content.mddocs/guides/runtime-adapter-switching.mddocs/guides/server-tools.mddocs/guides/streaming.mddocs/guides/structured-outputs.mddocs/guides/tool-approval.mddocs/guides/tool-architecture.mddocs/guides/tools.mddocs/guides/tree-shaking.mddocs/protocol/chunk-definitions.mddocs/protocol/http-stream-protocol.mddocs/protocol/sse-protocol.mddocs/reference/functions/chat.mddocs/reference/functions/combineStrategies.mddocs/reference/functions/maxIterations.mddocs/reference/functions/streamToText.mddocs/reference/functions/summarize.mddocs/reference/functions/toHttpResponse.mddocs/reference/functions/toHttpStream.mddocs/reference/functions/toServerSentEventsResponse.mddocs/reference/functions/untilFinishReason.mddocs/reference/interfaces/SummarizeAdapter.mddocs/reference/interfaces/TextAdapter.md
🚧 Files skipped from review as they are similar to previous changes (11)
- docs/reference/functions/toHttpStream.md
- docs/guides/agentic-cycle.md
- docs/reference/functions/streamToText.md
- docs/guides/runtime-adapter-switching.md
- docs/guides/tool-architecture.md
- docs/reference/functions/untilFinishReason.md
- docs/reference/functions/combineStrategies.md
- docs/getting-started/overview.md
- docs/guides/client-tools.md
- docs/reference/functions/toServerSentEventsResponse.md
- docs/reference/functions/maxIterations.md
🧰 Additional context used
🧠 Learnings (7)
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Implement framework integrations using the headless `tanstack/ai-client` for state management with framework-specific hooks (useChat) on top
Applied to files:
docs/guides/streaming.mddocs/adapters/openai.mddocs/guides/migration.mddocs/guides/tree-shaking.mddocs/guides/tools.mddocs/guides/multimodal-content.mddocs/api/ai.mddocs/reference/functions/chat.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to **/*.{ts,tsx} : Implement stream processing with StreamProcessor for handling chunked responses and support partial JSON parsing for streaming AI responses
Applied to files:
docs/adapters/openai.mddocs/guides/migration.mddocs/getting-started/quick-start.mddocs/reference/functions/toHttpResponse.mddocs/api/ai.mddocs/protocol/sse-protocol.mddocs/protocol/chunk-definitions.mddocs/guides/server-tools.mddocs/reference/functions/chat.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to packages/typescript/*/src/adapters/*.ts : Create individual adapter implementations for each provider capability (text, embed, summarize, image) with separate exports to enable tree-shaking
Applied to files:
docs/guides/migration.mddocs/guides/tree-shaking.mddocs/guides/tools.mddocs/guides/multimodal-content.mddocs/reference/interfaces/TextAdapter.mddocs/reference/functions/summarize.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to **/*.{ts,tsx} : Use tree-shakeable adapter architecture for provider implementations - export specialized adapters (text, embedding, summarize, image) as separate imports from `/adapters` subpath rather than monolithic adapters
Applied to files:
docs/guides/migration.mddocs/guides/tree-shaking.mddocs/reference/interfaces/TextAdapter.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to packages/typescript/*/src/index.ts : Export tree-shakeable adapters with clear subpath exports in package.json (e.g., `tanstack/ai/adapters`, `tanstack/ai-openai/adapters`) to minimize bundle size
Applied to files:
docs/guides/migration.mddocs/guides/tree-shaking.md
📚 Learning: 2025-12-27T21:39:29.563Z
Learnt from: harry-whorlow
Repo: TanStack/ai PR: 117
File: packages/typescript/ai-ollama/src/meta/model-meta-llama-guard3.ts:70-75
Timestamp: 2025-12-27T21:39:29.563Z
Learning: The standard template comments in ai-ollama model-meta files (like "Models with text, image, audio, video (no document)") should not be modified to match individual model capabilities, as they are intentionally kept uniform across all model-meta-*.ts files for consistency, regardless of what each specific model actually supports.
Applied to files:
docs/guides/multimodal-content.md
📚 Learning: 2025-12-13T17:09:09.794Z
Learnt from: CR
Repo: TanStack/ai PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-13T17:09:09.794Z
Learning: Applies to **/*.{ts,tsx} : Use type-safe per-model configuration with provider options typed based on selected model to ensure compile-time safety
Applied to files:
docs/reference/interfaces/SummarizeAdapter.md
🪛 markdownlint-cli2 (0.18.1)
docs/protocol/sse-protocol.md
82-82: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
88-88: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
🔇 Additional comments (16)
docs/reference/functions/summarize.md (1)
45-45: LGTM! Model identifiers updated correctly.The documentation updates to use
gpt-5-miniare correct—this model is OpenAI's faster, more cost-efficient version designed for well-defined tasks like summarization. The changes appropriately update all four code examples in this reference documentation.Also applies to: 54-54, 63-63, 71-71
docs/reference/functions/toHttpResponse.md (1)
45-45: Same verification needed: confirm gpt-5.2 model availability.This example also references the
gpt-5.2model. The verification request from the previous file applies here as well.docs/guides/tools.md (1)
191-191: Same verification needed: confirm gpt-5.2 model availability.Both examples reference the
gpt-5.2model. The verification request from the first file applies to these lines as well.Also applies to: 283-283
docs/guides/tool-approval.md (1)
68-68: The gpt-5.2 model is valid and available. OpenAI announced GPT-5.2 on December 11, 2025, and it is publicly available via the OpenAI API under the aliasesgpt-5.2,gpt-5.2-chat-latest, andgpt-5.2-pro. The documentation example is accurate.docs/reference/functions/chat.md (1)
53-53: Model reference is valid. The gpt-5.2 model is available from OpenAI (announced December 11, 2025) and is correctly referenced in the examples across lines 53, 65, 74, and 85.docs/reference/interfaces/SummarizeAdapter.md (1)
16-16: Verify consistency with PR objectives regarding model selection for this example.Both
gpt-5.2andgpt-5-miniare valid, available OpenAI models as of January 2026. However, confirm whether the PR objectives specify a particular model for summarization examples and whethergpt-5.2aligns with those objectives.docs/reference/interfaces/TextAdapter.md (1)
16-16: Documentation example update is consistent.The model name in the generic parameter example has been updated to match the broader documentation changes. Ensure
gpt-5.2is a valid model identifier as noted in other files.docs/guides/multimodal-content.md (1)
61-61: Model name updates are consistent with multimodal examples.The examples have been updated to use
gpt-5.2, which should support the multimodal content demonstrated (text + image). Verification of model capabilities is requested in the modality support section above.Also applies to: 285-285, 287-287
docs/guides/structured-outputs.md (1)
54-54: No issue found.gpt-5.2is a valid OpenAI model with full support for Structured Outputs, including schema-constrained JSON responses and function-calling. The documentation reference is accurate.Likely an incorrect or invalid review comment.
docs/guides/tree-shaking.md (1)
46-46: No action needed. The OpenAI modelsgpt-5.2andgpt-5-miniare valid and publicly available (announced December 11, 2025). The documentation examples are correct.Likely an incorrect or invalid review comment.
docs/adapters/openai.md (1)
22-22: Consistent with PR objectives; model validity verification needed.The updates correctly replace
gpt-4owithgpt-5.2in chat examples andgpt-4o-miniwithgpt-5-miniin the summarization example (line 144). Specialized models for audio/transcription (whisper-1,tts-1) and image generation (gpt-image-1) are appropriately left unchanged, which aligns with the PR description.The model identifier validity checks requested in previous files will confirm whether these updates are correct.
Also applies to: 38-38, 66-66, 95-95, 107-107, 144-144
docs/api/ai.md (1)
24-24: Bothgpt-5.2andgpt-5-miniare valid OpenAI model identifiers confirmed as available in the API as of January 2026. These model references in the documentation are accurate and will not cause integration failures.docs/protocol/chunk-definitions.md (1)
19-19: No action required —gpt-5.2is a valid and currently available OpenAI model identifier.OpenAI released GPT-5.2 on December 11, 2025, with the API identifier "gpt-5.2" available to developers. The documentation's use of this model identifier is accurate and reflects currently available APIs.
docs/protocol/http-stream-protocol.md (1)
86-88: LGTM! Valid model identifier updates.The documentation examples have been correctly updated to use the newer
gpt-5.2model identifier, which is confirmed as a valid OpenAI model released in December 2025. All changes are documentation-only and maintain consistency throughout the file.Also applies to: 94-95, 101-101, 133-137, 153-153, 184-184, 239-239, 363-365
docs/protocol/sse-protocol.md (1)
77-77: LGTM! Valid model identifier updates.The SSE protocol documentation has been correctly updated to use the
gpt-5.2model identifier throughout all examples. The changes are consistent and maintain the integrity of the protocol documentation.Also applies to: 83-83, 89-89, 123-127, 149-149, 177-177, 226-226, 307-311
docs/guides/migration.md (1)
33-33: LGTM! Migration guide updated with current model identifiers.The migration guide has been appropriately updated to use
gpt-5.2throughout all code examples. This ensures developers following the migration guide will be using current, valid model identifiers when implementing the new API patterns.Also applies to: 45-45, 52-52, 126-126, 141-141, 160-160, 174-174, 200-200, 214-214, 231-231, 233-233, 238-238, 260-260, 280-280, 374-374, 401-401
| - `gpt-5.2`, `gpt-5-mini`: text, image | ||
| - `gpt-5.2-audio-preview`: text, image, audio |
There was a problem hiding this comment.
🧩 Analysis chain
🌐 Web query:
OpenAI gpt-5.2 gpt-5-mini gpt-5.2-audio-preview model capabilities modalities support
💡 Result:
Summary — current OpenAI docs (Jan 2026):
-
GPT-5.2 (API name gpt-5.2 / gpt-5.2-chat-latest / gpt-5.2-pro): text I/O; image input only; audio not supported; video not supported. Supports Responses, Chat Completions, Realtime, Assistants, Batch, streaming, function-calling, tools (web search, file search, image generation, etc.), structured outputs; Pro/Thinking/Instant variants and reasoning-effort settings. [1][3]
-
GPT-5 mini (gpt-5-mini): smaller/cost‑optimized GPT‑5 family model — text I/O; image input only; audio not supported; video not supported. Same core endpoints and many features (streaming, function calling, code interpreter support), but lower cost and limits. [2]
-
“gpt-5.2-audio-preview”: there is no public GPT‑5.2 audio-preview model documented — OpenAI’s audio-preview model listings (e.g., gpt-4o-audio-preview) are separate and show audio-capable models under distinct names. If you need audio I/O, use the documented audio-preview models (check OpenAI’s audio model pages) or confirm whether OpenAI has released an explicit GPT‑5 audio preview. [1][4]
Sources:
[1] OpenAI Platform — GPT‑5.2 model page.
[2] OpenAI Platform — GPT‑5 mini model page.
[3] OpenAI — Introducing GPT‑5.2 announcement.
[4] OpenAI Platform — example audio-preview model (gpt‑4o‑audio‑preview) for reference.
Remove or replace the non-existent gpt-5.2-audio-preview model reference.
According to OpenAI's official documentation, gpt-5.2 and gpt-5-mini correctly support text and image modalities. However, gpt-5.2-audio-preview does not exist in OpenAI's public documentation. If audio support is needed, use OpenAI's documented audio-preview models (e.g., gpt-4o-audio-preview) instead, or remove this entry if not yet available.
🤖 Prompt for AI Agents
In @docs/guides/multimodal-content.md around lines 106 - 107, The docs list a
non-existent model `gpt-5.2-audio-preview`; update the multimodal table so it
only references valid models: keep `gpt-5.2` and `gpt-5-mini` as text+image, and
either remove the `gpt-5.2-audio-preview` row entirely or replace it with a
documented audio-capable model such as `gpt-4o-audio-preview` (or another
official audio-preview model) and adjust the modalities column accordingly.
2 participants
Summary
gpt-5.2instead ofgpt-4ogpt-5-miniinstead ofgpt-4o-minigpt-4o-audio-preview,whisper-1, etc.)The Story Behind This PR 📖
You know how AI coding agents have knowledge cutoffs? Well, at their cutoff,
gpt-4owas the latest and greatest OpenAI model. This means every AI agent out there defaults to suggestinggpt-4oin their code examples.I've had to fight AI agents so many times to stop using
gpt-4oin my code that now whenever I seegpt-4oin documentation, it triggers me. 😅So I decided to update all the examples to
gpt-5.2- not just for my sanity, but also because:gpt-4oTest Plan
Summary by CodeRabbit
✏️ Tip: You can customize this high-level summary in your review settings.