GPT-5.4 mini (medium)

OpenAI·Released 2026-03-17

Compare

Quality Index

37.7

57th of 440

Top 13%

Coding Index

37.5

36th of 350

Top 10%

Price/1M

$1.69

530th cheapest

445% above median

Top 79%

Speed

270 tok/s

Top 3%

TTFT

9.14s

Market Position

GPT-5.4 mini (medium)Market Average

Pricing

Input

$0.75

per 1M tokens

Output

$4.50

per 1M tokens

Blended

$1.69

per 1M tokens

Cheaper than 21% of models. Median price is $0.31/1M tokens.

Cost Calculator

Tokens per day1M

100K100M

Daily

$1.69

Monthly

$50.64

vs. Similar Models

Step 3.5 FlashQ:+0.1

$0.15-91%

GPT-5.4 nano (medium)Q:+0.4

$0.46-73%

Kimi K2.5 (Non-reasoning)Q:-0.4

$1.20-29%

Qwen3.5 27B (Non-reasoning)Q:-0.5

$0.82-51%

Performance

270

tokens/sec

Faster than 97% of models

9.14

seconds

Faster than 8% of models

Market Median

46 tok/s

490% faster

Median TTFT

0.43s

2025% slower

Throughput/Dollar

160

tok/s per $/1M

Speed Comparison

Gemini 2.5 Flash-Lite (Reasoning)

271 tok/s+0%

gpt-oss-120B (low)

273 tok/s+1%

gpt-oss-120B (high)

277 tok/s+3%

Benchmarks

MMLU-ProNot evaluated

GPQA Diamond

82.3%

HLE

17.1%

LiveCodeBenchNot evaluated

SciCode

44.2%

TerminalBench Hard

34.1%

MATH-500Not evaluated

AIMENot evaluated

AIME 2025Not evaluated

IFBench

64.8%

Long Context Recall

61.3%

Tau2

36.5%

Market AverageTop Score

Quick Compare

Similar Models

Step 3.5 Flash

StepFun

Q: 37.8$0.15/1M256K ctx

Slower: 66%Cheaper: 91%

GPT-5.4 nano (medium)

OpenAI

Q: 38.1$0.46/1M

Slower: 27%Cheaper: 73%

Kimi K2.5 (Non-reasoning)

Kimi

Q: 37.3$1.20/1M

Slower: 87%Cheaper: 29%

Qwen3.5 27B (Non-reasoning)

Alibaba

Q: 37.2$0.82/1M

Slower: 66%Cheaper: 51%

Qwen3.5 35B A3B (Reasoning)

Alibaba

Q: 37.1$0.69/1M262K ctx

Slower: 59%Cheaper: 59%

Claude 4.5 Haiku (Reasoning)

Anthropic

Q: 37.1$2.00/1M

Slower: 59%Pricier: 18%

Compare all 7 models