Which is better for coding — GPT-5 Medium or GPT-5.4?

Direct comparison on the SWE-Bench benchmark is not available. We recommend reviewing other metrics on the comparison page.

Which model is cheaper — GPT-5 Medium or GPT-5.4?

GPT-5 Medium is cheaper for input: $0.75 per 1M tokens vs $2.50.

Which has a larger context window — GPT-5 Medium or GPT-5.4?

GPT-5.4 supports a larger context: 1,000,000 tokens vs 400,000.

GPT-5 Medium vs GPT-5.4: Specs & Benchmark Comparison

Characteristic	GPT-5 Medium	GPT-5.4
Company	OpenAI	OpenAI
Release Date	August 6, 2025	March 5, 2026
Parameters	—	—
Multimodal	Yes	Yes
Context (input)	400K	1.0M
Context (output)	64K	128K
Input Price / 1M	$0.75	$2.50
Output Price / 1M	$3.00	$15.00
Average Score	0.9	0.9
Benchmarks
GPQA	0.9	0.9

Visual Benchmark Comparison

GPT-5 Medium

GPT-5.4

GPQA0.9 vs 0.9

0.9

Verdict

GPT-5.4 leads in 2 out of 4 comparison categories.

Overall Performance

Both models show comparable average scores: GPT-5 Medium — 0.9, GPT-5.4 — 0.9.

API Cost

GPT-5 Medium is 4.7x cheaper: input $0.75/1M vs $2.50/1M tokens.

Context Window

GPT-5.4 supports a larger context: 1M vs 400K tokens.

Recency

GPT-5.4 is newer: released 3/5/2026 vs 8/6/2025.

More About These Models

GPT-5 Medium

OpenAI — specs, benchmarks, API

GPT-5.4

OpenAI — specs, benchmarks, API

Related Comparisons

GPT-5 Medium vs GPT-5.1 Instant GPT-5 Medium vs GPT-5.1 Thinking GPT-5 High vs GPT-5 Medium GPT-5 Medium vs GPT-5.1 High GPT-5 Medium vs GPT-5.4 mini GPT-5.1 High vs GPT-5.4

All model comparisons →

The GPT-5 Medium and GPT-5.4 comparison is updated for 2026. Data includes benchmark results, API pricing, context window size and other specifications. For more detailed information, visit the GPT-5 Medium or GPT-5.4 page. See also the complete list of AI model comparisons.