Leaderboard

FrontierMathofficial site ↗

8 models scored · metric: accuracy. Prices are the cheapest listed across all serving providers. Cost/run and Pts/$ are priced for the heavy reasoning token profile — thinking tokens outnumber the answer several times over.

A benchmark is priced for the workload it exercises, not a fixed input:output blend: an agent loop and a single hard question bill very differently on the same rate card. Profile: 2k in · 1.5k out · 8k reasoning · 0% cache hit · 12k context.

Scores are facts as publicly reported by labs (see report links below and on each model page) and aggregated via models.dev. FrontierMath is maintained by its own project — Model Pulse is not affiliated with or endorsed by it.

Best value · top 5 by pts per dollar · heavy reasoning profile

1.GPT-5.6 Luna21,623
2.GPT-5.54,673
3.GPT-5.6 Terra3,859
4.GPT-5.6 Sol899
5.GPT-5.4814

#
🥇	GPT-6 Astraopenai	97.6	$10	$50	$0.495	197	1.05M
🥈	GPT-5.6 Solopenai	89	$2	$10	$0.099	899	1.05M
🥉	GPT-5.6 Terraopenai	84.9	$1.50	$2	$0.022	3,859	1.05M
4	GPT-5.6 Lunaopenai	78.6	$0.06	$0.37	0.363¢	21,623	1.05M
5	GPT-5.5 Proopenai	52.4	$27.27	$163.64	$1.609	32.6	1.05M
6	GPT-5.5openai	51.7	$0.188	$1.13	$0.011	4,673	1.05M
7	GPT-5.4 Proopenai	50	$27	$160	$1.574	31.8	1.05M
8	GPT-5.4openai	47.6	$0.75	$6	$0.059	814	1.05M

Score reports: openai.com · openai.com · openai.com — full source URLs are linked on each model page.

Cheapest scorer: GPT-5.6 Luna at $0.06 input /M.