Skip to content

ministral-3b-2512

Mistral

average±4.4 (95%): 51.0
leaderboard rankof 114: ≈103
benchmarksof 12: 12
per 1,000 documentsmean over its benchmarks: $0.42
seconds per inputmean response time: 8.2

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

By task

Data correction

39.2

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

37.3

Best of 107 models with all its benchmarks: gpt-6-astra, 97.0.
Information extraction

61.1

Best of 107 models with all its benchmarks: grok-4.6, 90.8.
Transcription

56.0

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 Dataother prompt	66.7±33.8	≈103of 115	$0.78$0.004 run	23.44	T1956 · 29 Oct 2026prompt_v2.txt
Blacklist Cards	71.9±5.2	≈106of 114	$0.30$0.010 run	1.98	T1963 · 24 Oct 2026prompt.txt
Book Advert XML Repairother prompt	1.9±3.4	≈105of 119	$0.16$0.008 run	5.05	T1965 · 24 Oct 2026prompt_v2.txt
Business Letters	41.0±5.7	≈107of 114	$0.45$0.026 run	1.98	T1957 · 24 Oct 2026prompt.txt
Company Listsother prompt	69.2±14.1	≈84of 114	$0.35$0.005 run	4.64	T1964 · 24 Oct 2026prompt_v2.txt
Duty Rosters	21.3±8.8	≈101of 107	$0.83$0.017 run	29.89	T1969 · 24 Oct 2026prompt.txt
Fraktur Adverts	44.6±27.4	≈94of 114	$0.40$0.002 run	8.34	T1960 · 29 Oct 2026prompt_optimized.txt
General Meeting Minutes	75.7±10.7	≈85of 108	$0.38$0.003 run	5.67	T1968 · 24 Oct 2026prompt.txt
Library Cardsother prompt	82.0±1.5	≈81of 114	$0.22$0.058 run	1.62	T1961 · 24 Oct 2026prompt_year-string.txt
Magazine Pagesother prompt	14.1±10.0	≈66of 115	$0.27$0.012 run	2.28	T1967 · 24 Oct 2026prompt_explicit-json.txt
Medieval Manuscripts	46.6±17.2	≈95of 114	$0.39$0.005 run	6.41	T1962 · 29 Oct 2026prompt.txt
Personnel Cards	76.6±7.1	≈94of 114	$0.52$0.032 run	7.0	T1966 · 24 Oct 2026prompt.txt

Price

From	Input, $ per million tokens	Output, $ per million tokens	Source
24 Oct 2026	0.1	0.1	archived price page

Every run