How much does the OpenAI GPT-5.4 API cost?

GPT-5.4 API pricing is $2.50 per million input tokens and $15.00 per million output tokens. Use our calculator at aiapicost.com for exact cost estimates based on your usage.

Which AI model is cheapest for API usage?

The cheapest AI API models change frequently. Use aiapicost.com to compare real-time pricing across 400+ models from OpenAI, Anthropic, Google, DeepSeek, and more. DeepSeek and open-source models typically offer the lowest per-token costs.

How do AI API token costs work?

AI APIs charge per token (roughly 0.75 words). Costs are split into input tokens (what you send) and output tokens (what the model generates). Output tokens are typically 2-5x more expensive. Prices are quoted per 1 million tokens.

Claude vs ChatGPT: which is better?

Both are top-tier models. Claude excels at coding and instruction-following, while GPT-5.4 offers broader multimodal capabilities. Compare them head-to-head at aiapicost.com/compare with real benchmark data.

Live data from Artificial Analysis API

AI Model Benchmarks

Compare 510+ AI models across 12 benchmarks — Intelligence, Coding, Math, Science, and more. Data updated hourly.

Benchmarks:

510 models · click headers to sort

Model

Speed

$/1M

AA Index▼

GPQA

MMLU-Pro

LiveCode

AIME

HLE

GPT-5.5 (xhigh)

OpenAI

63 t/s

$11.3

60.2

93.5%

—

44.3%

GPT-5.5 (high)

OpenAI

61 t/s

$11.3

58.9

93.2%

—

43.0%

Claude Opus 4.7 (Adaptive Reasoning, Max Effort)

Anthropic

64 t/s

$10.9

57.3

91.4%

—

39.6%

Gemini 3.1 Pro Preview

Google

133 t/s

$4.5

57.2

94.1%

—

44.7%

GPT-5.4 (xhigh)

OpenAI

80 t/s

$5.6

56.8

92.0%

—

41.6%

GPT-5.5 (medium)

OpenAI

61 t/s

$11.3

56.7

92.6%

—

40.6%

Kimi K2.6

Kimi

40 t/s

$1.7

53.9

91.1%

—

35.9%

MiMo-V2.5-Pro

Xiaomi

55 t/s

$1.5

53.8

86.6%

—

33.8%

GPT-5.3 Codex (xhigh)

OpenAI

80 t/s

$4.8

53.6

91.5%

—

39.9%

Grok 4.3

xAI

86 t/s

$1.6

53.2

90.1%

—

35.0%

Claude Opus 4.6 (Adaptive Reasoning, Max Effort)

Anthropic

49 t/s

$10.9

52.9

89.6%

—

36.7%

Muse Spark

Meta

—

$0.00

52.2

88.4%

—

39.9%

Claude Opus 4.7 (Non-reasoning, High Effort)

Anthropic

62 t/s

$10.9

51.8

88.5%

—

31.2%

Qwen3.6 Max Preview

Alibaba

36 t/s

$2.9

51.8

88.8%

—

28.9%

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Anthropic

62 t/s

$6.6

51.7

87.5%

—

30.0%

DeepSeek V4 Pro (Reasoning, Max Effort)

DeepSeek

31 t/s

$2.2

51.5

88.8%

—

35.9%

GLM-5.1 (Reasoning)

Z AI

53 t/s

$2.1

51.4

86.8%

—

28.0%

GPT-5.2 (xhigh)

OpenAI

69 t/s

$4.8

51.3

90.3%

87.4%

88.9%

99.0%

35.4%

GPT-5.5 (low)

OpenAI

58 t/s

$11.3

50.8

91.0%

—

31.0%

Qwen3.6 Plus

Alibaba

52 t/s

$1.1

50.0

88.2%

—

25.7%

DeepSeek V4 Pro (Reasoning, High Effort)

DeepSeek

30 t/s

$2.2

49.8

90.5%

—

33.5%

GLM-5 (Reasoning)

Z AI

76 t/s

$1.6

49.8

82.0%

—

27.2%

Claude Opus 4.5 (Reasoning)

Anthropic

62 t/s

$10.9

49.7

86.6%

89.5%

87.1%

91.3%

28.4%

MiniMax-M2.7

MiniMax

51 t/s

$0.53

49.6

87.4%

—

28.1%

Grok 4.20 0309 v2 (Reasoning)

xAI

90 t/s

$3.0

49.3

91.1%

—

32.2%

MiMo-V2-Pro

Xiaomi

67 t/s

$1.5

49.2

87.0%

—

28.3%

MiMo-V2.5

Xiaomi

88 t/s

$0.72

49.0

84.9%

—

25.2%

GPT-5.2 Codex (xhigh)

OpenAI

94 t/s

$4.8

49.0

89.9%

—

33.5%

GPT-5.4 mini (xhigh)

OpenAI

178 t/s

$1.7

48.9

87.5%

—

26.6%

Grok 4.20 0309 (Reasoning)

xAI

90 t/s

$3.0

48.5

88.5%

—

30.0%

Gemini 3 Pro Preview (high)

Google

125 t/s

$4.5

48.4

90.8%

89.8%

91.7%

95.7%

37.2%

GPT-5.4 (low)

OpenAI

63 t/s

$5.6

47.9

87.1%

—

28.9%

GPT-5.1 (high)

OpenAI

118 t/s

$3.4

47.7

87.3%

87.0%

86.8%

94.0%

26.5%

GLM-5-Turbo

Z AI

—

$0.00

46.8

84.7%

—

25.4%

Kimi K2.5 (Reasoning)

Kimi

49 t/s

$1.1

46.8

87.9%

—

29.4%

GPT-5.2 (medium)

OpenAI

—

$4.8

46.6

86.4%

85.9%

89.4%

96.7%

24.9%

DeepSeek V4 Flash (Reasoning, Max Effort)

DeepSeek

71 t/s

$0.17

46.5

89.4%

—

32.1%

Claude Opus 4.6 (Non-reasoning, High Effort)

Anthropic

44 t/s

$10.9

46.5

84.0%

—

18.6%

Gemini 3 Flash Preview (Reasoning)

Google

197 t/s

$1.1

46.4

89.8%

89.0%

90.8%

97.0%

34.7%

DeepSeek V4 Flash (Reasoning, High Effort)

DeepSeek

—

$0.17

46.0

86.7%

—

27.8%

Qwen3.6 27B (Reasoning)

Alibaba

64 t/s

$1.4

45.8

84.2%

—

21.6%

Qwen3.5 397B A17B (Reasoning)

Alibaba

52 t/s

$1.4

45.0

89.3%

—

27.3%

MiMo-V2-Omni-0327

Xiaomi

107 t/s

$0.80

44.9

85.5%

—

20.4%

GPT-5 Codex (high)

OpenAI

171 t/s

$3.4

44.6

83.7%

86.5%

84.0%

98.7%

25.6%

GPT-5 (high)

OpenAI

77 t/s

$3.4

44.6

85.4%

87.1%

84.6%

94.3%

26.5%

Claude Sonnet 4.6 (Non-reasoning, High Effort)

Anthropic

49 t/s

$6.6

44.4

79.9%

—

13.2%

GPT-5.4 nano (xhigh)

OpenAI

147 t/s

$0.46

44.0

81.7%

—

26.5%

KAT Coder Pro V2

KwaiKAT

114 t/s

$0.53

43.8

85.5%

—

16.0%

GLM-5.1 (Non-reasoning)

Z AI

42 t/s

$2.1

43.8

83.9%

—

25.6%

Qwen3.6 35B A3B (Reasoning)

Alibaba

183 t/s

$0.56

43.5

84.1%

—

20.2%

MiMo-V2-Omni

Xiaomi

105 t/s

$0.00

43.4

82.8%

—

19.9%

GPT-5.1 Codex (high)

OpenAI

183 t/s

$3.4

43.1

86.0%

84.9%

95.7%

23.4%

Claude Opus 4.5 (Non-reasoning)

Anthropic

51 t/s

$10.9

43.1

81.0%

88.9%

73.8%

62.7%

12.9%

Claude 4.5 Sonnet (Reasoning)

Anthropic

47 t/s

$6.6

43.0

83.4%

87.5%

71.4%

88.0%

17.3%

Kimi K2.6 (Non-reasoning)

Kimi

37 t/s

$1.7

42.9

78.8%

—

18.2%

GLM 5V Turbo (Reasoning)

Z AI

—

$0.00

42.9

80.9%

—

15.8%

Claude Sonnet 4.6 (Non-reasoning, Low Effort)

Anthropic

52 t/s

$6.6

42.6

79.7%

—

10.8%

GLM-4.7 (Reasoning)

Z AI

107 t/s

$1.0

42.1

85.9%

85.6%

89.4%

95.0%

25.1%

Qwen3.5 27B (Reasoning)

Alibaba

90 t/s

$0.82

42.1

85.8%

—

22.2%

GPT-5 (medium)

OpenAI

76 t/s

$3.4

42.0

84.2%

86.7%

70.3%

91.7%

23.5%

Claude 4.1 Opus (Reasoning)

Anthropic

37 t/s

$32.8

42.0

80.9%

88.0%

65.4%

80.3%

11.9%

Hy3-preview (Reasoning)

Tencent

159 t/s

$0.00

41.9

86.7%

—

25.5%

MiniMax-M2.5

MiniMax

89 t/s

$0.53

41.9

84.8%

—

19.1%

DeepSeek V3.2 (Reasoning)

DeepSeek

—

$0.34

41.7

84.0%

86.2%

92.0%

22.2%

Qwen3.5 122B A10B (Reasoning)

Alibaba

158 t/s

$1.1

41.6

85.7%

—

23.4%

MiMo-V2-Flash (Feb 2026)

Xiaomi

137 t/s

$0.15

41.5

83.5%

—

20.0%

Grok 4

xAI

44 t/s

$8.5

41.5

87.7%

86.6%

81.9%

92.7%

23.9%

Gemini 3 Pro Preview (low)

Google

—

$4.5

41.3

88.7%

89.5%

85.7%

86.7%

27.6%

GPT-5 mini (high)

OpenAI

86 t/s

$0.69

41.2

82.8%

83.7%

83.8%

90.7%

19.7%

GPT-5.5 (Non-reasoning)

OpenAI

57 t/s

$11.3

40.9

76.8%

—

12.6%

Kimi K2 Thinking

Kimi

114 t/s

$1.1

40.9

83.8%

84.8%

85.3%

94.7%

22.3%

o3-pro

OpenAI

26 t/s

$35.0

40.7

84.5%

—

GLM-5 (Non-reasoning)

Z AI

67 t/s

$1.6

40.6

66.6%

—

7.2%

Qwen3.5 397B A17B (Non-reasoning)

Alibaba

52 t/s

$1.4

40.1

86.1%

—

18.8%

Qwen3 Max Thinking

Alibaba

49 t/s

$2.4

39.8

86.1%

—

26.2%

MiniMax-M2.1

MiniMax

87 t/s

$0.53

39.4

83.0%

87.5%

81.0%

82.7%

22.2%

DeepSeek V4 Pro (Non-reasoning)

DeepSeek

31 t/s

$2.2

39.3

71.7%

—

7.7%

Gemma 4 31B (Reasoning)

Google

35 t/s

$0.00

39.2

85.7%

—

22.7%

Mistral Medium 3.5

Mistral

153 t/s

$3.0

39.2

74.8%

—

12.8%

GPT-5 (low)

OpenAI

72 t/s

$3.4

39.2

80.8%

86.0%

76.3%

83.0%

18.4%

MiMo-V2-Flash (Reasoning)

Xiaomi

142 t/s

$0.15

39.2

84.6%

84.3%

86.8%

96.3%

21.1%

Claude 4 Opus (Reasoning)

Anthropic

40 t/s

$32.8

39.0

79.6%

87.3%

63.6%

73.3%

11.7%

GPT-5 mini (medium)

OpenAI

77 t/s

$0.69

38.9

80.3%

82.8%

69.2%

85.0%

14.6%

Claude 4 Sonnet (Reasoning)

Anthropic

50 t/s

$6.6

38.7

77.7%

84.2%

65.5%

74.3%

9.6%

Grok 4.1 Fast (Reasoning)

xAI

98 t/s

$0.28

38.6

85.3%

85.4%

82.2%

89.3%

17.6%

Qwen3.5 Omni Plus

Alibaba

56 t/s

$1.5

38.6

82.6%

—

13.9%

GPT-5.1 Codex mini (high)

OpenAI

176 t/s

$0.69

38.6

81.3%

82.0%

83.6%

91.7%

16.9%

Step 3.5 Flash 2603

StepFun

164 t/s

$0.00

38.5

82.6%

—

22.6%

OpenAI

95 t/s

$3.5

38.4

82.7%

85.3%

80.8%

88.3%

20.0%

GPT-5.4 nano (medium)

OpenAI

156 t/s

$0.46

38.1

76.1%

—

14.7%

Step 3.5 Flash

StepFun

154 t/s

$0.15

37.8

83.1%

—

19.1%

GPT-5.4 mini (medium)

OpenAI

172 t/s

$1.7

37.7

82.3%

—

17.1%

Kimi K2.5 (Non-reasoning)

Kimi

59 t/s

$1.2

37.3

78.9%

—

12.3%

Qwen3.5 27B (Non-reasoning)

Alibaba

89 t/s

$0.83

37.2

84.2%

—

13.2%

Claude 4.5 Haiku (Reasoning)

Anthropic

139 t/s

$2.2

37.1

67.2%

76.0%

61.5%

83.7%

9.7%

Qwen3.6 27B (Non-reasoning)

Alibaba

64 t/s

$1.4

37.1

82.9%

—

13.6%

Claude 4.5 Sonnet (Non-reasoning)

Anthropic

45 t/s

$6.6

37.1

72.7%

86.0%

59.0%

37.0%

7.1%

Qwen3.5 35B A3B (Reasoning)

Alibaba

119 t/s

$0.69

37.1

84.5%

—

19.7%

DeepSeek V4 Flash (Non-reasoning)

DeepSeek

72 t/s

$0.17

36.5

71.6%

—

7.0%

100

JT-35B-Flash

China Mobile

—

$0.00

36.1

82.9%

—

6.1%

Showing top 100 of 510 models. Use search/filter to narrow down.

Benchmark Guide

Intelligence

Source ↗

Composite score across math, science, coding

GPQA

Source ↗

Graduate-level science Q&A (Diamond)

MMLU-Pro

Source ↗

Knowledge & reasoning across 57 subjects

LiveCodeBench

Source ↗

Live coding benchmark with new problems

AIME 2025

Source ↗

American Invitational Math Exam

MATH-500

Source ↗

Competition-level math problems

HLE

Source ↗

Humanity's Last Exam - hardest questions

Coding

Source ↗

Composite coding benchmark score

Math

Source ↗

Composite math benchmark score

SciCode

Source ↗

Scientific coding problems

IFBench

Source ↗

Instruction following benchmark

TerminalBench

Source ↗

Terminal/CLI task completion

Compare pricing for all models side by side

Open AI API Cost Calculator →

AI Model Benchmarks

Benchmark Guide

Tools

Guides

Comparisons