Skip to main content
Radar

Five or fewer confirmed AI changes, with original sources, on mornings when something changed.A free source-linked morning brief.

See the free Radar Brief
BenchLM
OpenAI logo
Model A
GPT-5.2-Codex

OpenAI

57.78/100

Supported · Public rank #95

90% interval 54.8–60.8

GPT-5.2-Codex vs GPT-5.4

Updated September 3, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

OpenAI logo
Model B
GPT-5.4

OpenAI

72.89/100

Supported · Public rank #20

90% interval 69.1–76.7

Decision reading

GPT-5.4 has the higher public score, 72.89 versus 57.78, and the 90% score intervals do not overlap.

3 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.

Which one for your work

Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.

  • Long documents

    Prompts that approach the documented context limit

    GPT-5.4

    GPT-5.4 has the larger documented context window.

    Confidence: documented

  • Chat turn cost

    1K fresh input + 500 output tokens

    GPT-5.2-Codex

    GPT-5.2-Codex has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.

    Confidence: listed-rates

  • Cache-heavy agent loop cost

    200K cached + 20K fresh input + 10K output tokens

    GPT-5.4

    GPT-5.4 has the lower estimated token cost for this stated workload. GPT-5.2-Codex has no published cached-input rate, so cached tokens use its listed input rate.

    Confidence: rate-fallback

Show secondary and unsupported calls

  • Repository review cost

    50K fresh input + 3K output tokens

    GPT-5.2-Codex

    GPT-5.2-Codex has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.

    Confidence: listed-rates

  • Coding work

    Code generation, repair, and software-engineering tasks

    Not enough matched evidence

    No shared weighted benchmark basis supports a winner.

    Confidence: limited

  • Agentic work

    Tool use, computer use, and multi-step task completion

    Not enough matched evidence

    No shared weighted benchmark basis supports a winner.

    Confidence: limited

What is actually comparable

Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.

Shared results
3
GPT-5.2-Codex only
0
GPT-5.4 only
34
Like-for-like categories
0 / 8

Category results, on a stated basis

Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.

Agentic

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
77.2
Weighted basis
0 vs 3 rows
Reading
Not comparable

Coding

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
57.7
Weighted basis
0 vs 1 rows
Reading
Not comparable

Reasoning

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
74.0
Weighted basis
0 vs 1 rows
Reading
Not comparable

Knowledge

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
57.6
Weighted basis
0 vs 2 rows
Reading
Not comparable

Math

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
42.5
Weighted basis
0 vs 2 rows
Reading
Not comparable

Multilingual

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Multimodal

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
73.2
Weighted basis
0 vs 3 rows
Reading
Not comparable

Instruction following

Not comparable
GPT-5.2-Codex
Not measured
GPT-5.4
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Shape of the matched evidence

Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.

A shared-evidence shape is not available.

BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.

What each workload costs

Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.

Chat turn

1K fresh input + 500 output tokens

GPT-5.2-Codex
$0.00875
Fits in one request
GPT-5.4
$0.01
Fits in one request

GPT-5.2-Codex has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

GPT-5.2-Codex
$0.1295
Fits in one request
GPT-5.4
$0.17
Fits in one request

GPT-5.2-Codex has the lower modeled cost

Costs use the listed standard API rates.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

GPT-5.2-Codex
$0.525
Fits in one request
Cached input priced at the published list-input rate
GPT-5.4
$0.25
Fits in one request

GPT-5.4 has the lower modeled cost

GPT-5.2-Codex has no published cached-input rate, so cached tokens use its listed input rate.

Specification differences

Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.

Context window

Maximum documented context; output-token limits may be lower.

GPT-5.2-Codex

400K

GPT-5.4

1.05M

OpenAI pricing

Cached-input rate

A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.

GPT-5.2-Codex

Not published

GPT-5.4

$0.25 per 1M cached input tokens

OpenAI pricing

Documented inputs

GPT-5.2-Codex

Not sourced

GPT-5.4

Not sourced

Documented outputs

GPT-5.2-Codex

Not sourced

GPT-5.4

Not sourced

Provider availability

GPT-5.2-Codex

Not sourced

GPT-5.4

Not sourced

Reasoning profile

GPT-5.2-Codex

Reasoning

GPT-5.4

Reasoning

Weight access

GPT-5.2-Codex

Proprietary

GPT-5.4

Proprietary

License

GPT-5.2-Codex

Proprietary

GPT-5.4

Proprietary

Release date

GPT-5.2-Codex

2025-12-18

GPT-5.4

2026-03-05

If you already use one of these models
Deployment change
Both entries list OpenAI as the provider. Confirm endpoint, model ID, limits, and feature support before switching.
Quality signal
GPT-5.4 has the higher public score, 72.89 versus 57.78, and the 90% score intervals do not overlap.
Workload cost
Repository review: $0.1295 vs $0.17. Cache-heavy agent loop: $0.525 vs $0.25.
Context tradeoff
GPT-5.4 has the larger documented window (1.05M).

Run the same representative tasks against both endpoints before changing production traffic.

Benchmark evidence

The full public result ledger is available for audit without forcing a wide desktop table onto a phone.

Browse raw public benchmark evidence37 rows

Agentic

  • Gert Labs

    Shared source
    GPT-5.2-Codex51.79%
    GPT-5.464.89%

    GPT-5.4 leads this result

  • JobBench

    Shared source
    GPT-5.2-Codex26.0%
    GPT-5.438.9%

    GPT-5.4 leads this result

  • Terminal-Bench 2.0

    GPT-5.2-Codex—
    GPT-5.475.1%
    Source

    Not directly comparable

  • CyberGym

    GPT-5.2-Codex—
    GPT-5.479.0%
    Source

    Not directly comparable

  • BrowseComp

    GPT-5.2-Codex—
    GPT-5.482.7%
    Source

    Not directly comparable

  • OSWorld-Verified

    GPT-5.2-Codex—
    GPT-5.475%
    Source

    Not directly comparable

  • MCP Atlas

    GPT-5.2-Codex—
    GPT-5.470.6%
    Source

    Not directly comparable

  • Toolathlon

    GPT-5.2-Codex—
    GPT-5.454.6%
    Source

    Not directly comparable

  • τ²-bench results

    GPT-5.2-Codex—
    GPT-5.498.9%
    Source

    Not directly comparable

  • Claw-Eval

    GPT-5.2-Codex—
    GPT-5.460.3%
    Source

    Not directly comparable

  • DeepSearchQA

    GPT-5.2-Codex—
    GPT-5.473.6%
    Source

    Not directly comparable

  • ResearchClawBench

    GPT-5.2-Codex—
    GPT-5.415.3%
    Source

    Not directly comparable

  • ExploitGym

    GPT-5.2-Codex—
    GPT-5.46.0%
    Source

    Not directly comparable

Coding

  • Vibe Code Bench

    Shared source
    GPT-5.2-Codex37.91%
    GPT-5.467.42%

    GPT-5.4 leads this result

  • LiveCodeBench Pro

    GPT-5.2-Codex—
    GPT-5.487.5%
    Source

    Not directly comparable

  • SWE-bench Pro

    GPT-5.2-Codex—
    GPT-5.457.7%
    Source

    Not directly comparable

  • React Native Evals

    GPT-5.2-Codex—
    GPT-5.485.3%
    Source

    Not directly comparable

Reasoning

  • ARC-AGI-2

    GPT-5.2-Codex—
    GPT-5.474.0%
    Source

    Not directly comparable

  • ARC-AGI-3

    GPT-5.2-Codex—
    GPT-5.40.2%
    Source

    Not directly comparable

Knowledge

  • GPQA

    GPT-5.2-Codex—
    GPT-5.492.8%
    Source

    Not directly comparable

  • HLE

    GPT-5.2-Codex—
    GPT-5.452.1%
    Source

    Not directly comparable

  • HLE w/o tools

    GPT-5.2-Codex—
    GPT-5.439.8%
    Source

    Not directly comparable

  • GPQA-D

    GPT-5.2-Codex—
    GPT-5.492.8%
    Source

    Not directly comparable

  • HealthBench Hard

    GPT-5.2-Codex—
    GPT-5.440.1%
    Source

    Not directly comparable

  • MedXpertQA (Text)

    GPT-5.2-Codex—
    GPT-5.459.6%
    Source

    Not directly comparable

  • HealthBench Professional

    GPT-5.2-Codex—
    GPT-5.448.1%
    Source

    Not directly comparable

Math

  • FrontierMath v2 (Tiers 1-3)

    GPT-5.2-Codex—
    GPT-5.447.600%
    Source

    Not directly comparable

  • FrontierMath v2 (Tier 4)

    GPT-5.2-Codex—
    GPT-5.427.100%
    Source

    Not directly comparable

Multimodal

  • MMMU-Pro

    GPT-5.2-Codex—
    GPT-5.481.2%
    Source

    Not directly comparable

  • OfficeQA Pro

    GPT-5.2-Codex—
    GPT-5.453.2%
    Source

    Not directly comparable

  • MMMU-Pro w/ Python

    GPT-5.2-Codex—
    GPT-5.482.1%
    Source

    Not directly comparable

  • CharXiv

    GPT-5.2-Codex—
    GPT-5.482.8%
    Source

    Not directly comparable

  • ERQA

    GPT-5.2-Codex—
    GPT-5.465.4%
    Source

    Not directly comparable

  • SimpleVQA

    GPT-5.2-Codex—
    GPT-5.461.1%
    Source

    Not directly comparable

  • ScreenSpot Pro

    GPT-5.2-Codex—
    GPT-5.485.4%
    Source

    Not directly comparable

  • ZeroBench

    GPT-5.2-Codex—
    GPT-5.441.0%
    Source

    Not directly comparable

  • MedXpertQA (MM)

    GPT-5.2-Codex—
    GPT-5.477.1%
    Source

    Not directly comparable

Frequently asked questions

Which is better, GPT-5.2-Codex or GPT-5.4?

GPT-5.4 has the higher public score, 72.89 versus 57.78, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.

Which is better for coding, GPT-5.2-Codex or GPT-5.4?

The published evidence does not provide a shared weighted coding basis for both models, so BenchLM does not name a coding winner.

Which is better for agentic tasks, GPT-5.2-Codex or GPT-5.4?

The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.

Which costs less, GPT-5.2-Codex or GPT-5.4?

For the stated presets, chat costs $0.00875 on GPT-5.2-Codex and $0.01 on GPT-5.4; repository review costs $0.1295 and $0.17; the cache-heavy agent loop costs $0.525 and $0.25. GPT-5.2-Codex has no published cached-input rate, so cached tokens use its listed input rate.

Which has the larger context window, GPT-5.2-Codex or GPT-5.4?

GPT-5.4 has the larger documented context window: 1.05M, compared with 400K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated September 3, 2026

Watch GPT-5.2-Codex vs GPT-5.4

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 2,000+ readers.