Data points: 126
Model compare
GPT-5.3 Codex vs GPT-5.4
The readout for GPT-5.3 Codex and GPT-5.4, before the detailed comparison sheet.
Key Takeaways
Weighted outcome: GPT-5.4. Benchmark capability categories carry 80%, while price, API performance, and availability carry 20%.
Decision read
GPT-5.4
GPT-5.4 has the higher weighted result; Model A / B score 34.7 to 65.3.
Evidence depth
126 data points
Includes 31 benchmark rows, 0 audit samples, and 10 provider examples.
Selection signal
Start with GPT-5.4
The charts below split 41 high-signal samples across speed, scores, and audit health.
Change comparison
Switch either side of this report to compare another model with the same LMSpeed data pipeline.
Select a different model to open a new comparison URL.
Comparison sheet
This report only uses LMSpeed data for GPT-5.3 Codex and GPT-5.4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.
| Model compare | GPT-5.3 Codex | GPT-5.4 |
|---|---|---|
| Overall leader | Contender | Leading |
| Weighted overall score | 34.7 pts | 65.3 pts |
| Benchmark category leads | 2 categories | 4 categories |
| Operational advantages | Average speed, First-token latency | Cheapest input price, Free providers, Provider coverage |
| Context window | 400K tokens | 1.1M tokens |
| Max output | 128K tokens | 128K tokens |
| Modalities | Input TextImageFile Output Text | Input TextImageFile Output Text |
| Features | Text inputImage inputFile inputText outputTool callingStructured outputsJSON modeReasoning | Text inputImage inputFile inputText outputTool callingStructured outputsJSON modeReasoning |
The overall result weights benchmark capability categories at 80% and price, API speed/latency, and availability at 20%. Recent test volume does not affect the winner, and missing benchmark categories are excluded.
Model metadata
Model metadata
| Model compare | GPT-5.3 Codex | GPT-5.4 |
|---|---|---|
| Developer | OpenAI | OpenAI |
| Released | Feb 2026 | Mar 2026 |
| Parameters | No data | No data |
| Tokenizer | GPT | GPT |
| Knowledge cutoff | No data | No data |
| OpenRouter ID | openai/gpt-5.3-codex | openai/gpt-5.4 |
| References | No data | No data |
When to choose each model
This report only uses LMSpeed data for GPT-5.3 Codex and GPT-5.4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.
GPT-5.3 Codex
GPT-5.3 Codex is stronger in benchmark categories (Reasoning, Instruction following) and operational dimensions (Average speed, First-token latency).
GPT-5.4
GPT-5.4 is stronger in benchmark categories (Agents, Coding, Knowledge, Multimodal) and operational dimensions (Cheapest input price, Free providers, Provider coverage).
Benchmark score comparison
Third-party benchmark profile synced into LMSpeed; only metrics available for both models are shown.
Category performance
Compare benchmark category scores on a 0-100 scale. Select a category to inspect the gap.
- Model A coverage
- 6 / 8
- Model B coverage
- 7 / 8
- Shared
- 6 shared categories
Avg. score
GPT-5.3 Codex
55.3
Avg. score
GPT-5.4
56.6
Agents
GPT-5.4 leads by 4.2
Coding
GPT-5.4 leads by 5.1
Reasoning
GPT-5.3 Codex leads by 4.1
Knowledge
GPT-5.4 leads by 1.4
Math
GPT-5.4
Multilingual
No data
Multimodal
GPT-5.4 leads by 1.7
Instruction following
GPT-5.3 Codex leads by 0.6
Professional benchmark details
Metric-level scores with benchmark source, rank depth, confidence, error, and evaluation date where available.
Professional benchmark details
Metric-level scores with benchmark source, rank depth, confidence, error, and evaluation date where available.
GPT-5.3 Codex
winner68.0
Rank #26/111 · confidence 3 · eval date 2026-02-05
GPT-5.4
63.0
Rank #40/111 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner$1.75/M
Rank #140/186 · confidence 4
GPT-5.4
$2.50/M
Rank #154/186 · confidence 4
GPT-5.3 Codex
winner$4.81/M
Rank #155/186 · confidence 4
GPT-5.4
$5.63/M
Rank #158/186 · confidence 4
GPT-5.3 Codex
winner$14.00/M
Rank #156/186 · confidence 4
GPT-5.4
$15.00/M
Rank #159/186 · confidence 4
GPT-5.3 Codex
winner78.3
Rank #5/67 · confidence 3 · eval date 2026-02-05
GPT-5.4
57.2
Rank #51/67 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner83.3
Rank #6/106 · confidence 3 · eval date 2026-02-05
GPT-5.4
82.0
Rank #14/106 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
16.9
Rank #22/88 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner23.4
Rank #11/88 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner91.5%
Rank #21/218 · confidence 4
GPT-5.4
74.8%
Rank #122/218 · confidence 4
GPT-5.3 Codex
winner42.5%
Rank #22/216 · confidence 4
GPT-5.4
11.3%
Rank #116/216 · confidence 4
GPT-5.3 Codex
61.8
Rank #4/33 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner67.4
Rank #3/33 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
53.2
Rank #26/113 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner56.6
Rank #10/113 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
56.8
Rank #26/50 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner57.7
Rank #21/50 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner52.0
Rank #50/87 · confidence 3 · eval date 2026-02-05
GPT-5.4
42.8
Rank #69/87 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner52.9
Rank #11/106 · confidence 3 · eval date 2026-02-05
GPT-5.4
50.8
Rank #15/106 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
89.2
Rank #21/106 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner91.7
Rank #13/106 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
42.5
Rank #21/114 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner43.7
Rank #15/114 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner79.5
Rank #19/83 · confidence 3 · eval date 2026-02-05
GPT-5.4
74.3
Rank #30/83 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
91.5
Rank #23/113 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner92.0
Rank #21/113 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner10.9
Rank #25/45 · confidence 3 · eval date 2026-02-05
GPT-5.4
5.8
Rank #28/45 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
32.5
Rank #43/117 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner39.0
Rank #28/117 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner75.6
Rank #20/56 · confidence 3 · eval date 2026-02-05
GPT-5.4
69.3
Rank #28/56 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner78.5
Rank #22/68 · confidence 3 · eval date 2026-02-05
GPT-5.4
78.4
Rank #23/68 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
1177.0
Rank #59/78 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner1234.0
Rank #42/78 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner92.6
Rank #10/52 · confidence 3 · eval date 2026-02-05
GPT-5.4
90.6
Rank #16/52 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner75.4
Rank #18/84 · confidence 3 · eval date 2026-02-05
GPT-5.4
73.9
Rank #21/84 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
86.0
Rank #42/82 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner98.9
Rank #2/82 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
57.5
Rank #13/49 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner64.9
Rank #4/49 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
33.7
Rank #11/24 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner38.9
Rank #7/24 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
64.7
Rank #21/27 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner75.0
Rank #11/27 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
winner77.3
Rank #11/53 · confidence 3 · eval date 2026-02-05
GPT-5.4
75.1
Rank #13/53 · confidence 4 · eval date 2026-03-05
GPT-5.3 Codex
62.8
Rank #37/76 · confidence 3 · eval date 2026-02-05
GPT-5.4
winner72.0
Rank #28/76 · confidence 4 · eval date 2026-03-05
API audit comparison
Latest completed audits from shared providers, with four safety and integrity score groups plus report links.
API audit comparison
Latest completed audits from shared providers, with four safety and integrity score groups plus report links.
| Provider | GPT-5.3 Codex | GPT-5.4 |
|---|---|---|
| No completed audits are available from shared providers yet. | ||
Provider examples
Speed aggregates and input/output pricing share each provider row for real API selection and migration cost checks.
Provider examples
Speed aggregates and input/output pricing share each provider row for real API selection and migration cost checks.
| Provider | GPT-5.3 Codex | GPT-5.4 |
|---|---|---|
Mars HK125 tests | GPT-5.3 Codex speed / latency N/A / N/A input / output No data | GPT-5.4 speed / latency 49 tok/s / 6084ms input / output No data |
Mars HK114 tests | GPT-5.3 Codex speed / latency 252 tok/s / 2196ms input / output No data | GPT-5.4 speed / latency 57 tok/s / 3556ms input / output No data |
星见雅 API55 tests | GPT-5.3 Codex speed / latency 38 tok/s / 4553ms input / output No data | GPT-5.4 speed / latency 49 tok/s / 5435ms input / output No data |
天宫造物50 tests | GPT-5.3 Codex speed / latency N/A / N/A input / output No data | GPT-5.4 speed / latency 50 tok/s / 7305ms input / output No data |
QuicklyAPI35 tests | GPT-5.3 Codex speed / latency N/A / N/A input / output No data | GPT-5.4 speed / latency 50 tok/s / 2467ms input / output No data |
GPT-5.3 Codex gpt-5.3-codex-spark speed / latency No data input / output $0.263/M/$2.10/M | GPT-5.4 gpt-5.4-2026-03-05 speed / latency No data input / output $0/M/$0/M | |
GPT-5.3 Codex gpt-5.3-codex speed / latency No data input / output $1.34/M/$10.74/M | GPT-5.4 gpt-5.4 speed / latency No data input / output $0/M/$0/M | |
GPT-5.3 Codex gpt-5.3-codex-xhigh speed / latency No data input / output $0/request | GPT-5.4 gpt-5.4-high speed / latency No data input / output $0/request | |
GPT-5.3 Codex gpt-5.3-codex speed / latency No data input / output $0.0068/request | GPT-5.4 gpt-5.4 speed / latency No data input / output $0.014/request | |
GPT-5.3 Codex gpt-5.3-codex-spark speed / latency No data input / output $0.0073/request | GPT-5.4 gpt-5.4 speed / latency No data input / output $0.073/request |
FAQ
Weighted outcome: GPT-5.4. Benchmark capability categories carry 80%, while price, API performance, and availability carry 20%.
- Why is this comparison indexable?
- It has 6 verifiable comparison points, and both models have pricing or benchmark data.
- Are missing metrics invented?
- No. Metrics without LMSpeed data are omitted from this report.