Skip to main content
BenchLM
Advertise

Model comparison

GPT-5.3 Codex vs GPT-5.4 nano

Updated July 30, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

GPT-5.3 Codex

OpenAI

65.8/100

Supported · Public rank #30

90% interval 62.3–69.2

GPT-5.4 nano

OpenAI

66.0/100

Supported · Public rank #28

90% interval 55.5–76.4

GPT-5.4 nano has the higher public score estimate, 65.98 versus 65.75, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.

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.

  • Agentic work

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

    GPT-5.3 Codex

    GPT-5.3 Codex leads on the same 2 weighted benchmark rows.

    Confidence: limited

  • Chat turn cost

    1K fresh input + 500 output tokens

    GPT-5.4 nano

    GPT-5.4 nano 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 nano

    GPT-5.4 nano has the lower estimated token cost for this stated workload. GPT-5.3 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-5.4 nano

    GPT-5.4 nano 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

  • Long documents

    Prompts that approach the documented context limit

    No clear pick

    The documented context windows are equal.

    Confidence: documented

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-5.3 Codex only
5
GPT-5.4 nano only
10
Like-for-like categories
1 / 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

Like-for-like
GPT-5.3 Codex
71.4
GPT-5.4 nano
42.9
Weighted basis
2 vs 2 rows
Reading
GPT-5.3 Codex leads

Coding

Not comparable
GPT-5.3 Codex
67.2
GPT-5.4 nano
Not measured
Weighted basis
3 vs 0 rows
Reading
Not comparable

Reasoning

Not comparable
GPT-5.3 Codex
Not measured
GPT-5.4 nano
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Knowledge

Not comparable
GPT-5.3 Codex
Not measured
GPT-5.4 nano
43.8
Weighted basis
0 vs 2 rows
Reading
Not comparable

Math

Not comparable
GPT-5.3 Codex
Not measured
GPT-5.4 nano
21.0
Weighted basis
0 vs 2 rows
Reading
Not comparable

Multilingual

Not comparable
GPT-5.3 Codex
Not measured
GPT-5.4 nano
Not measured
Weighted basis
0 vs 0 rows
Reading
Not comparable

Multimodal

Not comparable
GPT-5.3 Codex
Not measured
GPT-5.4 nano
66.1
Weighted basis
0 vs 1 rows
Reading
Not comparable

Instruction following

Not comparable
GPT-5.3 Codex
Not measured
GPT-5.4 nano
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.

Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.

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-5.3 Codex
$0.00875
Fits in one request
GPT-5.4 nano
$0.00082
Fits in one request

GPT-5.4 nano has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

GPT-5.3 Codex
$0.1295
Fits in one request
GPT-5.4 nano
$0.01375
Fits in one request

GPT-5.4 nano 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-5.3 Codex
$0.525
Fits in one request
Cached input priced at the published list-input rate
GPT-5.4 nano
$0.0205
Fits in one request

GPT-5.4 nano has the lower modeled cost

GPT-5.3 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.

Cached-input rate

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

GPT-5.3 Codex

Not published

GPT-5.4 nano

$0.02 per 1M cached input tokens

OpenAI pricing

Reasoning profile

GPT-5.3 Codex

Reasoning

GPT-5.4 nano

Reasoning

Weight access

GPT-5.3 Codex

Proprietary

GPT-5.4 nano

Proprietary

License

GPT-5.3 Codex

Proprietary

GPT-5.4 nano

Proprietary

Release date

GPT-5.3 Codex

2026-02-05

GPT-5.4 nano

2026-03-17

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 nano has the higher public score estimate, 65.98 versus 65.75, but the 90% score intervals overlap.
Workload cost
Repository review: $0.1295 vs $0.01375. Cache-heavy agent loop: $0.525 vs $0.0205.
Context tradeoff
Both models list 400K.

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 evidence18 rows

Agentic

  • Terminal-Bench 2.0

    GPT-5.3 Codex77.3%
    Source
    GPT-5.4 nano46.3%
    Source

    GPT-5.3 Codex leads this result

  • OSWorld-Verified

    GPT-5.3 Codex64.7%
    Source
    GPT-5.4 nano39%
    Source

    GPT-5.3 Codex leads this result

  • Gert Labs

    GPT-5.3 Codex57.47%
    Source
    GPT-5.4 nano—

    Not directly comparable

  • JobBench

    GPT-5.3 Codex33.7%
    Source
    GPT-5.4 nano—

    Not directly comparable

  • MCP Atlas

    GPT-5.3 Codex—
    GPT-5.4 nano56.1%
    Source

    Not directly comparable

  • Toolathlon

    GPT-5.3 Codex—
    GPT-5.4 nano35.5%
    Source

    Not directly comparable

  • τ²-bench results

    GPT-5.3 Codex—
    GPT-5.4 nano92.5%
    Source

    Not directly comparable

Coding

  • SWE-bench Verified

    GPT-5.3 Codex85%
    Source
    GPT-5.4 nano—

    Not directly comparable

  • SWE-bench Pro

    GPT-5.3 Codex56.8%
    Source
    GPT-5.4 nano—

    Not directly comparable

  • SWE-Rebench

    GPT-5.3 Codex58.2%
    Source
    GPT-5.4 nano—

    Not directly comparable

  • Vibe Code Bench

    Shared source
    GPT-5.3 Codex61.77%
    GPT-5.4 nano26.10%

    GPT-5.3 Codex leads this result

Knowledge

  • GPQA

    GPT-5.3 Codex—
    GPT-5.4 nano82.8%
    Source

    Not directly comparable

  • HLE

    GPT-5.3 Codex—
    GPT-5.4 nano37.7%
    Source

    Not directly comparable

  • HLE w/o tools

    GPT-5.3 Codex—
    GPT-5.4 nano24.3%
    Source

    Not directly comparable

Math

  • FrontierMath v2 (Tiers 1-3)

    GPT-5.3 Codex—
    GPT-5.4 nano25.860%
    Source

    Not directly comparable

  • FrontierMath v2 (Tier 4)

    GPT-5.3 Codex—
    GPT-5.4 nano6.250%
    Source

    Not directly comparable

Multimodal

  • MMMU-Pro

    GPT-5.3 Codex—
    GPT-5.4 nano66.1%
    Source

    Not directly comparable

  • MMMU-Pro w/ Python

    GPT-5.3 Codex—
    GPT-5.4 nano69.5%
    Source

    Not directly comparable

Frequently asked questions

Which is better, GPT-5.3 Codex or GPT-5.4 nano?

GPT-5.4 nano has the higher public score estimate, 65.98 versus 65.75, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.

Which is better for coding, GPT-5.3 Codex or GPT-5.4 nano?

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-5.3 Codex or GPT-5.4 nano?

GPT-5.3 Codex leads the like-for-like agentic tasks comparison across 2 shared weighted benchmark rows.

Which costs less, GPT-5.3 Codex or GPT-5.4 nano?

For the stated presets, chat costs $0.00875 on GPT-5.3 Codex and $0.00082 on GPT-5.4 nano; repository review costs $0.1295 and $0.01375; the cache-heavy agent loop costs $0.525 and $0.0205. GPT-5.3 Codex has no published cached-input rate, so cached tokens use its listed input rate.

Which has the larger context window, GPT-5.3 Codex or GPT-5.4 nano?

Both models list the same context window, 400K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated July 30, 2026

Watch GPT-5.3 Codex vs GPT-5.4 nano

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

Read a sample issue

Join 2,000+ readers.