GPT-5.2 Pro vs GPT-5 mini

Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.

GPT-5.2 Pro is clearly ahead on the aggregate, 90 to 69. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

GPT-5.2 Pro's sharpest advantage is in coding, where it averages 84.8 against 42.8. The single biggest benchmark swing on the page is LiveCodeBench, 81 to 37.

GPT-5.2 Pro gives you the larger context window at 400K, compared with 128K for GPT-5 mini.

Quick Verdict

Pick GPT-5.2 Pro if you want the stronger benchmark profile. GPT-5 mini only becomes the better choice if its workflow or ecosystem matters more than the raw scoreboard.

Agentic

GPT-5.2 Pro

GPT-5.2 Pro

85.9

GPT-5 mini

65.7

88
Terminal-Bench 2.0
68
88
BrowseComp
70
82
OSWorld-Verified
60

Coding

GPT-5.2 Pro

GPT-5.2 Pro

84.8

GPT-5 mini

42.8

93
HumanEval
80
83
SWE-bench Verified
41
81
LiveCodeBench
37
89
SWE-bench Pro
49

Multimodal & Grounded

GPT-5.2 Pro

GPT-5.2 Pro

96

GPT-5 mini

83.8

96
MMMU-Pro
86
96
OfficeQA Pro
81

Reasoning

GPT-5.2 Pro

GPT-5.2 Pro

95.2

GPT-5 mini

81.8

97
SimpleQA
84
95
MuSR
82
98
BBH
87
93
LongBench v2
80
95
MRCRv2
79

Knowledge

GPT-5.2 Pro

GPT-5.2 Pro

81.5

GPT-5 mini

62.8

99
MMLU
88
99
GPQA
86
97
SuperGPQA
84
95
OpenBookQA
82
90
MMLU-Pro
73
44
HLE
16
93
FrontierScience
75

Instruction Following

GPT-5.2 Pro

GPT-5.2 Pro

95

GPT-5 mini

82

95
IFEval
82

Multilingual

GPT-5.2 Pro

GPT-5.2 Pro

93.4

GPT-5 mini

80.1

96
MGSM
82
92
MMLU-ProX
79

Mathematics

GPT-5.2 Pro

GPT-5.2 Pro

98.2

GPT-5 mini

87.2

99
AIME 2023
90
99
AIME 2024
92
99
AIME 2025
91
96
HMMT Feb 2023
86
98
HMMT Feb 2024
88
97
HMMT Feb 2025
87
97
BRUMO 2025
89
99
MATH-500
85

Frequently Asked Questions

Which is better, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro is ahead overall, 90 to 69. The biggest single separator in this matchup is LiveCodeBench, where the scores are 81 and 37.

Which is better for knowledge tasks, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for knowledge tasks in this comparison, averaging 81.5 versus 62.8. Inside this category, HLE is the benchmark that creates the most daylight between them.

Which is better for coding, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for coding in this comparison, averaging 84.8 versus 42.8. Inside this category, LiveCodeBench is the benchmark that creates the most daylight between them.

Which is better for math, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for math in this comparison, averaging 98.2 versus 87.2. Inside this category, MATH-500 is the benchmark that creates the most daylight between them.

Which is better for reasoning, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for reasoning in this comparison, averaging 95.2 versus 81.8. Inside this category, MRCRv2 is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for agentic tasks in this comparison, averaging 85.9 versus 65.7. Inside this category, OSWorld-Verified is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for multimodal and grounded tasks in this comparison, averaging 96 versus 83.8. Inside this category, OfficeQA Pro is the benchmark that creates the most daylight between them.

Which is better for instruction following, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for instruction following in this comparison, averaging 95 versus 82. Inside this category, IFEval is the benchmark that creates the most daylight between them.

Which is better for multilingual tasks, GPT-5.2 Pro or GPT-5 mini?

GPT-5.2 Pro has the edge for multilingual tasks in this comparison, averaging 93.4 versus 80.1. Inside this category, MGSM is the benchmark that creates the most daylight between them.

Last updated: March 12, 2026

Weekly LLM Benchmark Digest

Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.

Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.