OpenAI: GPT-5.2

Benchmarks gegenereerd uit Aibenchy-testsuites op : 19 feb 2026

Categorie-uitsplitsing

Categorie	Volledig geslaagde tests	Score	Consistentie	Slaagpercentage per poging	Instabiele tests	Redeneerscore	Kosten
Anti-AI Tricks	2/2	10.00	10.00	100.0%	0	8.25	$0.01131
Data parsing and extraction	2/2	10.00	10.00	100.0%	0	9.50	$0.02230
Domain specific	1/3	4.00	7.21	55.6%	1	3.17	$0.13697
Instructions following	2/2	9.50	10.00	100.0%	0	8.00	$0.01071
Puzzle Solving	2/3	8.00	10.00	66.7%	0	8.83	$0.02427

Vergelijk OpenAI: GPT-5.2 met...

#4 · Qwen

Redenering (medium)

Score: 8.42

Consistentie: 9.30

Slaagpercentage per poging: 86.1%

Instabiele tests: 1

Kosten per resultaat: 2.3151

Correcte tests: 10/12

Totale kosten: $0.23151

#6 · Anthropic

Redenering (medium)

Score: 7.00

Consistentie: 9.30

Slaagpercentage per poging: 69.4%

Instabiele tests: 1

Kosten per resultaat: 9.3797

Correcte tests: 8/12

Totale kosten: $0.75038

#3 · Google

Redenering (medium)

Score: 8.42

Consistentie: 10.00

Slaagpercentage per poging: 83.3%

Instabiele tests: 0

Kosten per resultaat: 0.8028

Correcte tests: 10/12

Totale kosten: $0.08029

Vergelijk OpenAI: GPT-5.2 met...