GPT-5.2 Codex API: Endpoint & Pricing
Deprecated 23 Jul 2026GPT-5.2-Codex is an OpenAI model name, but there is no public, reliable technical information available about this specific variant.
What is GPT-5.2-Codex?
GPT-5.2-Codex is a referenced OpenAI model name for which no official public specification or documentation is currently available. Because of this, concrete details about its capabilities, training data, or deployment context are not known. Its real-world use cases, performance characteristics, and positioning within OpenAI’s product lineup have not been formally described. Any relationship it may have to prior Codex or GPT model families has not been publicly clarified by OpenAI.
Can you self-host GPT-5.2-Codex?
No. We found no public weights for GPT-5.2-Codex on Hugging Face, so there is nothing to download, quantize or serve on your own GPUs: no VRAM budget, no vLLM or GGUF build to plan for. The only way to run it is over an API.
Skip the deploy — use LLM.API as your endpoint
No GPUs, no quantization trade-offs. Same model, OpenAI-compatible, up to 30% below list price.
Providers
Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick — Balanced (price + speed), Nitro (fastest), or Exacto (highest tool-calling accuracy).
List price by provider
InputOutputProvider list prices; the LLM.API discount applies on top.
| Provider | Input | Output | Cache read /M | Latency | Throughput | Uptime |
|---|---|---|---|---|---|---|
| in $1.75; out $14 per 1M tokens | 400K tokens | vision, tools, streaming, reasoning, web search, JSON, structured |
Prices and availability from the LLMAPI catalogue, updated nightly. Last updated 11 Oct 2026.
Try this model
Test GPT-5.2-Codex right here — free to start.
Suggestions for your first prompt
Code snippet
Call the model through the OpenAI-compatible API.
Switching from OpenAI? Change 2 lines.
- base_url="https://api.openai.com/v1"
- api_key="YOUR_OPENAI_KEY"
+ base_url="https://api.llmapi.ai/v1"
+ api_key="YOUR_LLMAPI_KEY"Everything else stays the same — SDK, message format, tools, streaming. Set model to gpt-5.2-codex and the call runs unchanged.
curl https://api.llmapi.ai/v1/chat/completions \
-H "Authorization: Bearer $LLMAPI_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-5.2-codex",
"messages": [
{"role": "user", "content": "Give me three crisp launch checklist items."}
],
"reasoning_effort": "none"
}'from openai import OpenAI
client = OpenAI(
api_key="YOUR_LLMAPI_KEY",
base_url="https://api.llmapi.ai/v1",
)
resp = client.chat.completions.create(
model="gpt-5.2-codex",
messages=[
{"role": "system", "content": "You are a precise product assistant."},
{"role": "user", "content": "Give me three crisp launch checklist items."},
],
reasoning_effort="none",
)
print(resp.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
apiKey: process.env.LLMAPI_KEY,
baseURL: "https://api.llmapi.ai/v1",
});
const resp = await client.chat.completions.create({
model: "gpt-5.2-codex",
messages: [
{ role: "user", content: "Give me three crisp launch checklist items." },
],
});
console.log(resp.choices[0].message.content);stream = client.chat.completions.create(
model="gpt-5.2-codex",
messages=[{"role": "user", "content": "Write a haiku about latency."}],
stream=True,
)
for chunk in stream:
delta = chunk.choices[0].delta.content
if delta:
print(delta, end="", flush=True){
"model": "gpt-5.2-codex",
"messages": [
{"role": "system", "content": "You are a precise product assistant."},
{"role": "user", "content": "Give me three crisp launch checklist items."}
],
"reasoning_effort": "none"
}ENDPOINTS
Supported endpoints & features
Base URL https://api.llmapi.ai/v1 · model ID gpt-5.2-codex · OpenAI-compatible.
POST /v1/chat/completionsChat Completions — supported
| Feature | Status | How to set it |
|---|---|---|
| Function calling (tools) | Supported | tools, tool_choice |
| Parallel tool calls | Not supported | parallel_tool_calls |
| JSON mode | Supported | response_format: {"type": "json_object"} |
| Structured outputs | Supported | response_format: {"type": "json_schema", …} |
| Streaming | Supported | stream: true |
| Vision (image input) | Supported | image_url content parts |
| Web search | Supported | web_search |
| Reasoning effort | none · low · medium · high · xhigh | reasoning_effort: "none" |
ERRORS
Errors & fallback routing
| Code | What it means | What to do |
|---|---|---|
401 | Missing or invalid API key. | Send Authorization: Bearer YOUR_LLMAPI_KEY and check the key is active. |
429 | Rate limited, or the upstream provider is throttling the request. | Back off and retry with jitter; the gateway also retries the request on another provider where one is available. |
5xx | Upstream provider error or timeout. | Retry; fallback routing sends the retry to the next healthy provider for this model. |
When a provider fails or throttles, the request is routed to the next provider serving this model — the ones listed in the providers table above.
Uptime
Last 30 days
26/30 days operational | 96.70% uptime
GPT-5.2 Codex in production
What the model is good at.
-
Conversational AI
Engages in multi-turn conversations, following instructions, maintaining context, and producing coherent, helpful responses across diverse domains.
-
Code Generation
Generates source code snippets or functions in various programming languages based on natural language specifications and problem descriptions.
-
Text Translation
Translates text between multiple languages, preserving meaning and tone while adapting to contextual nuances and idiomatic expressions.
-
Image Reasoning
Interprets images to answer questions or extract structured information, connecting visual content with textual instructions or prompts.
-
Visual Text Reading
Reads and interprets text appearing within images, such as documents, screenshots, or signs, enabling downstream understanding and processing.
What you get when you call GPT-5.2 Codex through LLM.API.
-
Unified AI Routing
Dynamically route each request to the optimal model across providers based on latency, cost, and quality—without changing your integration or redeploying.
One endpoint, any model -
Cost-Aware Orchestration
Automatically pick cheaper compatible models, enforce cost caps, and track spend per project so you can scale AI usage without runaway bills.
Minimize spend by default -
Resilient Fallbacks
Configure multi-provider fallbacks so requests seamlessly fail over on outages, throttling, or timeouts—no single vendor or region can take you down.
High availability by design -
End-to-End Observability
Inspect logs, latencies, costs, and provider errors for every call from a single dashboard, making it easy to debug issues and optimize performance.
See every token, everywhere -
Task-Level Abstractions
Call high-level tasks like chat, tools, or reranking instead of vendor-specific APIs, so you can swap models without rewriting business logic.
Code to tasks, not vendors -
High-Throughput Batch
Submit massive batches of requests with built-in rate control, retries, and progress tracking to efficiently process datasets, backfills, and offline workloads.
Process millions efficiently
When to Use — When NOT to Use
Use it if...
- You need a top-tier model for complex code generation across multiple programming languages.
- Your use case involves refactoring or modernizing large legacy codebases with minimal regressions.
- You need sophisticated bug localization and automatic patch suggestions for production-scale services.
- Your use case involves generating end-to-end applications, including backend, frontend, and tests.
- You need deep reasoning about code behavior, performance tradeoffs, and security implications.
- Your use case involves multi-file edits where the model must maintain architectural consistency.
- You need advanced assistance for API design, library authoring, and framework-level abstractions.
Avoid if...
- You need the absolute lowest-cost model for simple boilerplate or CRUD code.
- Your workload requires ultra-low-latency token streaming for high-frequency real-time interactions.
- You need strictly on-device or air-gapped deployment without relying on external cloud services.
- Your workload requires processing highly sensitive data where external hosted models are prohibited.
- You need a lightweight model for inexpensive bulk classification or simple text tagging tasks.
- Your workload requires strict deterministic outputs without any variability across generations or runs.
- You need a model specialized for long-form creative writing rather than code-centric reasoning.
COMMUNITY
What developers say about GPT-5.2-Codex
Summarised from public developer discussions and write-ups. Opinions are the sources’, not LLM.API’s.
- Developers on Hacker News praise GPT-5.2-Codex for its methodical approach to finding bugs and inconsistencies, stating its code quality is superior to competitors like Claude for complex code reviews.
- A GitHub contributor reports that the model can be overly concise and lack critical detail during the planning phase of agentic workflows, sometimes underperforming compared to the base GPT-5.2 model.
- Developers on GitHub and Hacker News describe the model as a 'workhorse' for professional tasks, though some note it can appear 'lazy' or 'indecisive' compared to non-Codex variants.
SOURCES
Frequently Asked Questions
-
What is GPT-5.2-Codex?
GPT-5.2-Codex is an OpenAI code-focused large language model optimized for software development, code generation, and complex debugging via LLM.API.
-
What is GPT-5.2-Codex best suited for?
GPT-5.2-Codex excels at generating, refactoring, and explaining code, handling multi-file repositories, and answering advanced programming and API design questions.
-
How is GPT-5.2-Codex priced when used through LLM.API?
LLM.API exposes GPT-5.2-Codex with usage-based pricing per input and output token; check your LLM.API dashboard or pricing docs for current rates.
-
What context window does GPT-5.2-Codex support on LLM.API?
GPT-5.2-Codex supports a large context window suitable for multi-file codebases; refer to LLM.API’s model table for the exact token limit.
-
How fast is GPT-5.2-Codex in terms of latency and throughput?
GPT-5.2-Codex typically responds with low latency and supports streaming, though actual speed depends on prompt size, output length, and LLM.API load.
-
What modalities does GPT-5.2-Codex support?
GPT-5.2-Codex supports text input and text code output; it is optimized for programming tasks rather than images or audio.
-
How do I call GPT-5.2-Codex via the LLM.API?
Use the LLM.API completion or chat endpoint with the model parameter set to "GPT-5.2-Codex" and authenticate using your LLM.API API key.
-
How does GPT-5.2-Codex compare to general-purpose GPT-5.2 models?
Compared to general GPT-5.2 variants, GPT-5.2-Codex is more capable on coding tasks but slightly less optimized for open-ended natural language generation.
-
What are the main limitations of GPT-5.2-Codex?
GPT-5.2-Codex can hallucinate incorrect code, lacks real-time access to your environment, and should not be trusted without tests, reviews, or security audits.
-
Can GPT-5.2-Codex access the internet or my private repositories through LLM.API?
No, GPT-5.2-Codex only sees data you include in the prompt or tool calls; it cannot independently browse or read private repositories.
COMPARE
Competitive Models
GPT-5.2
$1.75 in / $14 out per 1M tokens, 400,000 context — the closest current alternative to GPT-5.2 Codex.
Grok 4.6
$2 in / $6 out per 1M tokens, 500,000 context — the closest current alternative to GPT-5.2 Codex.
Kimi K2.7 Code HighSpeed
$1.9 in / $8 out per 1M tokens, 262,144 context — the closest current alternative to GPT-5.2 Codex.
GPT-5.2 Pro is an OpenAI frontier large language model optimized for strong general reasoning, coding, and multimodal assistant use in demanding, real-world applications.
GPT-5.1-Codex-Mini
Up to 30%GPT-5.1-Codex-Mini is an OpenAI code-focused model variant optimized for lightweight, fast software development assistance. It is notable for providing capable code generation and editing while using fewer resources than larger Codex-style models.
Text Embedding Ada 002
text-embedding-ada-002 is an OpenAI embedding model that converts text into numerical vectors for measuring semantic similarity. It is an improved, more performant successor to earlier Ada-based embedding models and became a widely used default for production applications.
Get one key to every model
Swap your API key. Keep your code.