GPT-5.2-Codex
OpenAI · released Jan 14, 2026
- Type
- Proprietary
- Input
- $1.75
- Output
- $14.00
- Cached
- $0.17
List price · per 1M tokens · OpenAI at 400K context · source ↗
Our take
Written Aug 3, 2026GPT-5.2-Codex is a hosted-only coding specialist from OpenAI that can handle up to 400,000 tokens in a single request. It scores well on raw coding tasks but its agentic-coding ability lags far behind, making it a strong reviewer rather than an autonomous builder.
Choose this for long-document code review or documentation tasks where 400,000 tokens of context matter. Use the standard tier when you do not need the premium one, or pick Azure if 12 tokens per second meets your throughput needs. Skip it if you need agentic coding workflows, self-hosting, or strong instruction-following scores.
The case for it
- Extremely long request limit for code and document processing: 400,000 tokens.
- Strong raw coding performance: LiveBench Coding 83.62 and Arena Code Elo 1338.4038.
- Highly stable scores over time, varying only 0.06 points across seven daily snapshots.
- Predictable premium pricing: the higher tier is exactly double the standard rate for both input and output.
The case against it
- Agentic coding falls 34.23 points below raw coding, scoring 41% lower on that LiveBench subtest.
- Instruction following is its weakest area at 66.45, well below its own overall average.
- Throughput is largely undisclosed: only Azure lists 12 tokens per second; no self-hosting option exists.
How good is it?
IntelligencePuzzles, maths, exam questions
GPT-5.2-Codex is not on Arena Text (overall), which is where the rating would come from, so there is no rating here. It is on LiveBench Data Analysis, in 8th of 35 with 78.2.
CodingWriting and fixing code on its own
GPT-5.2-Codex is not on Arena Coding, which is where the rating would come from, so there is no rating here. It is on Arena Code (WebDev), in 57th of 74 with 1338.4.
AgenticPlanning, calling tools, staying on task
GPT-5.2-Codex is not on Arena Agent (IPS), which is where the rating would come from, so there is no rating here. It is on SWE-bench Verified, in 7th of 39 with 72.8.
WritingWe do not rate this
Two boards come close and neither tests writing: Arena Creative Writing asks people which of two replies they prefer, and LiveBench Language tests whether a model understood a passage. So we show where GPT-5.2-Codex placed and give it no mark out of five.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done, which is why they get no rating.
Every published score for this model10 scoresEvery figure we hold, from 10 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Or rent it from someone else
Why this differs from the header. The strip above quotes OpenAI's own list price. This is the cheapest live offer at the widest standard context we hold, whoever is serving it — a reseller undercutting a lab is ordinary commerce, not an error.
- per 1M tokens
- $1.75 in / $14.00 out
- Context served
- 400K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouter | $1.75 / $14.00 | 400K | not measured | Unknown | Unknown | Unknown |
| Microsoft Azure AI | $1.75 / $14.00 | 400K | 13 tok/s | No | No | Confirmed |
| OpenAI | $1.75 / $14.00 | 400K | not measured | No | Yesunknown period | Unknown |
| OpenAIpriority tier | $3.50 / $28.00 | 400K | not measured | No | Yesunknown period | Unknown |
Microsoft Azure AIAcross the 4 listings we hold: 3 say they do not train on prompts, 0 say they do and 1 do not say. 1 appear in the zero-retention registry we check; the rest are unknown to us rather than confirmed either way.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
- ✓
- Supported
- ✗
- Not supported
- Not published
- host gave no parameter list
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouter | ✓ | ✓ | ✓ |
| Microsoft Azure AI | ✓ | ✓ | ✓ |
| OpenAI | ✓ | ✓ | ✓ |
| OpenAIpriority | ✓ | ✓ | ✓ |
Tool calling: 4 of 4 listings say yes. JSON output: 4 of 4 listings say yes. Strict schema: 4 of 4 listings say yes.
Models people weigh against GPT-5.2-Codex
When we formed this view
Dates behind this page
Prices last checked 9d ago
What we do not know about this model yet
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 4 listings do not say whether they train on prompts.
Licence and identifiers
What the licence allowsWe hold no licence record for this model. Inside are the identifiers you need to pull it — its Hugging Face repo where we have one, our slug and a machine-readable card.
Licence
Commercial API terms. We hold no licence record for this model, so there is nothing to summarise here.
Identifiers
- Modality record
- text+image->text
- Catalogue slug
- openai-gpt-5-2-codex