Claude Opus 4.1 vs GPT-5.2
Wins 0 of 8 areas
—
Wins 8 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Long documents · Following instructions
GPT-5.2 is the stronger all-rounder.
Scores updated · 30 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software04GPT-5.24 of 4 tests
- ReasoningHard problems that need careful thinking03GPT-5.23 of 3 tests
- FactsGetting facts right instead of making them up02GPT-5.22 of 2 tests
- MathCompetition and research-level math02GPT-5.22 of 2 tests
- AgentsCarrying out multi-step tasks on its own01GPT-5.21 of 1 test
- Images and chartsUnderstanding pictures, charts and video01GPT-5.21 of 1 test
- Long documentsFinding answers in very long texts01GPT-5.21 of 1 test
- Following instructionsDoing exactly what it is asked01GPT-5.21 of 1 test
Agents, images and charts, long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
GPT-5.2 costs 83% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where GPT-5.2 pulls ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+54.8points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+29.3points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+25.2points ahead
Where Claude Opus 4.1 pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 30 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGPT-5.2
- Terminal-Bench HardGPT-5.2 by 12.734.347+12.7
- SciCodeGPT-5.2 by 11.240.952.1+11.2
- SWE-bench VerifiedGPT-5.2 by 5.574.580+5.5
- LMArena · WebDevGPT-5.2 by 26 rating points13901416+26 rating
ReasoningGPT-5.2
- Humanity's Last ExamGPT-5.2 by 25.212.537.7+25.2
- CritPtGPT-5.2 by 11.6011.6+11.6
- GPQA DiamondGPT-5.2 by 9.480.990.3+9.4
FactsGPT-5.2
- SimpleQA VerifiedGPT-5.2 by 2.334.837.1+2.3
- Vectara HHEM hallucination ratelower is betterGPT-5.2 by 111.810.8+1
MathGPT-5.2
- FrontierMath Tiers 1-3 (v2)GPT-5.2 by 54.812.667.4+54.8
- FrontierMath Tier 4GPT-5.2 by 29.32.431.7+29.3
Other results15 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- AIME 2025GPT-5.2 by 2278100+22
- metr_re_benchGPT-5.2 by 201535+20
- metr_hcastGPT-5.2 by 16.460.677+16.4
- frontiermath_tier_4_v1GPT-5.2 by 14.64.218.8+14.6
- τ²-BenchGPT-5.2 by 1471.585.5+14
- τ²-Bench Telecom (AA run)GPT-5.2 by 13.471.484.8+13.4
- Artificial Analysis Coding IndexGPT-5.2 by 12.236.548.7+12.2
- AA IntelligenceGPT-5.2 by 7.622.830.4+7.6
- vectara_answer_rateGPT-5.2 by 7.692.4100+7.6
- Terminal-Bench 2.0GPT-5.2 by 7.546.554+7.5
- vectara_avg_summary_lengthGPT-5.2 by 57.2 rating points129.1186.3+57.2 rating
- τ²-Bench (Retail)Claude Opus 4.1 by 4.886.882+4.8
- vectara_factual_consistencyGPT-5.2 by 188.289.2+1
- metr_swaatie10099.6tie
- MMMLUtie89.589.6tie
Questions people ask
Which is better, Claude Opus 4.1 or GPT-5.2?
GPT-5.2 wins all eight areas we test: coding, agents, reasoning, facts, images and charts, math, long documents and following instructions. Claude Opus 4.1 wins none.
Which is better for coding?
GPT-5.2. It wins 4 of the 4 coding tests both models report; Claude Opus 4.1 wins none.
Which is cheaper?
Claude Opus 4.1 costs $15.00 per million input tokens and $75.00 per million output tokens; GPT-5.2 costs $1.75 and $14.00. That makes GPT-5.2 about 83% cheaper for the same work.
How do you compare the two?
We use the 30 benchmark tests both models have published scores on. The verdict counts the 15 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 15 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.