Skip to content

gemini-3-pro-preview

Google

legacy model Its test configurations are retired, so it is no longer run. The figures are from its best legacy runs, scored as they ran at the time; it is not ranked.

average±4.8 (95%): 79.4
benchmarksof 12: 9
per 1,000 documentsmean over its benchmarks: $82
seconds per inputmean response time: 101.3

Compare with other models Its row in the Matrix All its test runs

By task

Data correction

65.3

Best of 114 models with all its benchmarks: claude-opus-5-5, 82.7.
Transcription

86.1

Best of 114 models with all its benchmarks: gemini-4-argon, 96.2.

By benchmark

Benchmark	Score	Rank	Per 1,000 docs	s / input	Best run
Bibliographic Data	92.2±3.1	–	$101$0.50 run	77.35	T0415 · 24 Jan 2026prompt.txt
Blacklist Cards	95.7±2.4	–	$30$0.99 run	32.03	T0422 · 17 Feb 2026prompt.txt
Book Advert XML Repair	34.7±9.8	–	$152$7.58 run	117.03	T0445 · 17 Feb 2026prompt.txt
Business Letters	61.0±7.0	–	$51$2.93 run	47.76	T0416 · 25 Nov 2025prompt.txt
Company Lists	83.5±7.9	–	$53$0.80 run	74.28	T0423 · 17 Feb 2026prompt.txt
Fraktur Adverts	79.0±38.7	–	$120$0.60 run	93.15	T0419 · 25 Nov 2025prompt_optimized.txt
Library Cards	89.3±1.6	–	$29$7.51 run	35.34	T0420 · 25 Nov 2025prompt.txt
Medieval Manuscripts	83.5±12.7	–	$137$1.64 run	390.8	T0421 · 25 Nov 2025prompt.txt
Personnel Cards	95.9±1.3	–	$61$3.75 run	43.98	T0594 · 9 Feb 2026prompt.txt

Not run on: Duty Rosters, General Meeting Minutes, Magazine Pages.

Price

From	Input, $ per million tokens	Output, $ per million tokens	Source
24 Nov 2025	2.0	12.0	archived price page

Every run