vs
1:0
overall capability wins (1 ties across 2 scored). GPT-5.4 leads overall.
- Pricing3:0
- Specs1:0
Verdict
GPT-5.4 offers ~2.6x larger context (1.1M vs 400K); GPT-5.4 is ~5.0x cheaper on a blended token basis than GPT-5.6 Cyber.
- GPT-5.4 offers ~2.6x larger context (1.1M vs 400K)
- GPT-5.4 is ~5.0x cheaper on a blended token basis than GPT-5.6 Cyber
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- GPT-5.4
- $6.25
- GPT-5.6 Cyber
- $31.25
GPT-5.4 is about 5.0x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
OpenAI
OpenAI
Family
GPT-5.x
GPT-5.x
License
Proprietary
Proprietary
Open weights
No
No
Release
Mar 5, 2026
Aug 10, 2026
Knowledge cutoff
-
2026-02-16
API / provider
OpenAI
OpenAI
Modalities
text, image → text
text, image → text
Specs
Context window
1.1M
400K
Max output
128K
128K
Parameters
—
—
Pricing
Input $/1M
$2.50
$12.50
Output $/1M
$15
$75
Blended $/1M (3∶1)
$5.63
$28.13
Speed
tok/s
85
-
TTFT (s)
0.4
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
92.0
—
Humanity's Last Exam
43.7
—
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
50.8
—
AA-LCR v1.1
82.0
—
CritPt
23.4
—
MMMU-Pro
78.4
—
IFBench
73.9
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
57.7
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
78.3
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
56.6
—
CursorBench
—
—
SWE-Rebench
57.6
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,387
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,466
—
| Metric | GPT-5.4 | GPT-5.6 Cyber | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | OpenAI | — |
| Family | GPT-5.x | GPT-5.x | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Mar 5, 2026 | Aug 10, 2026 | — |
| Knowledge cutoff | - | 2026-02-16 | — |
| API / provider | OpenAI | OpenAI | — |
| Modalities | text, image → text | text, image → text | — |
| Specs | |||
| Context window | 1.1M | 400K | A +650K |
| Max output | 128K | 128K | tie |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $2.50 | $12.50 | A +$10 |
| Output $/1M | $15 | $75 | A +$60 |
| Blended $/1M (3∶1) | $5.63 | $28.13 | A +$22.50 |
| Speed | |||
| tok/s | 85 | - | — |
| TTFT (s) | 0.4 | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 92.0 | — | — |
| Humanity's Last Exam | 43.7 | — | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 50.8 | — | — |
| AA-LCR v1.1 | 82.0 | — | — |
| CritPt | 23.4 | — | — |
| MMMU-Pro | 78.4 | — | — |
| IFBench | 73.9 | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | 57.7 | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 78.3 | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 56.6 | — | — |
| CursorBench | — | — | — |
| SWE-Rebench | 57.6 | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,387 | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,466 | — | — |
GPT-5.4
- Context
- 1.1M / 128K out
- Parameters
- —
- Price
- $2.50 / $15
- Speed
- 85 tok/s · 0.4s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Solid mid-2026 OpenAI workhorse between GPT-5 and the 5.5/5.6 flagships.
GPT-5.6 Cyber
- Context
- 400K / 128K out
- Parameters
- —
- Price
- $12.50 / $75
- Speed
- —
- Modalities
- text, image → text
- License
- Proprietary
OpenAI's most capable and most permissive cybersecurity model (Daybreak program) for authorized vulnerability research and exploit validation — a GPT-5.6 Sol derivative at $12.5/$75; gated behind identity verification and legal declarations.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.
These two models have no overlapping published benchmarks in our dataset. Compare specs and pricing instead.