Compare/GPT-5.4 mini (medium) vs Claude 3.5 Haiku

GPT-5.4 mini (medium)vsClaude 3.5 Haiku

Side-by-side comparison of pricing, 12 benchmarks, and generation speed.

OpenAI

GPT-5.4 mini (medium)

Input
$0.75/M
Output
$4.5/M
Speed
200 tok/s
TTFT
3.77s
Anthropic

Claude 3.5 Haiku

Input
$0.8/M
Output
$4/M
Speed
TTFT

Winner by Category

Cheaper
Claude 3.5 Haiku
Faster (tok/s)
GPT-5.4 mini (medium)
Lower Latency
Claude 3.5 Haiku
Benchmarks (7-3)
GPT-5.4 mini (medium)

Pricing Comparison

MetricGPT-5.4 mini (medium)Claude 3.5 Haiku
Input ($/M tokens)$0.75$0.8
Output ($/M tokens)$4.5$4
Cost for 1M input + 100K output tokens:
GPT-5.4 mini (medium)$1.20
Claude 3.5 Haiku$1.20

Speed Comparison

Output Speed (tokens/s) — higher is better
GPT-5.4 mini (medium)
200 tok/s
Claude 3.5 Haiku
Time to First Token (seconds) — lower is better
GPT-5.4 mini (medium)
3.77s
Claude 3.5 Haiku

Benchmark Comparison

Data from Artificial Analysis API — 12 benchmarks

Intelligence Index
37.718.7
Coding Index
37.510.7
Math Index
GPQA Diamond
82.3%40.8%
MMLU-Pro
63.4%
LiveCodeBench
31.4%
AIME 2025
MATH-500
72.1%
Humanity's Last Exam
17.1%3.5%
SciCode
44.2%27.4%
IFBench
64.8%42.8%
TerminalBench
34.1%2.3%
GPT-5.4 mini (medium)7 wins
3 winsClaude 3.5 Haiku

Frequently Asked Questions

Which is cheaper, GPT-5.4 mini (medium) or Claude 3.5 Haiku?

Claude 3.5 Haiku is cheaper overall. Its blended price (3:1 input/output ratio) is $1.60/M tokens vs $1.69/M for GPT-5.4 mini (medium).

Which model performs better on benchmarks?

GPT-5.4 mini (medium) wins 7 out of 12 benchmarks compared to 3 for Claude 3.5 Haiku. See the detailed benchmark chart above for per-category results.

Which is faster for real-time applications?

GPT-5.4 mini (medium) generates tokens faster at 200 tok/s vs 0 tok/s. However, Claude 3.5 Haiku has lower time-to-first-token (0.00s vs 3.77s).

When should I use GPT-5.4 mini (medium) vs Claude 3.5 Haiku?

Choose based on your priorities: Claude 3.5 Haiku for lower cost, GPT-5.4 mini (medium) for stronger benchmark performance, and GPT-5.4 mini (medium) for faster generation. For latency-sensitive apps, check the TTFT comparison above.