Ministral 3 14B vs Mistral Large 3
Wins 0 of 7 areas
—
Wins 6 of 7 areas
Coding · Reasoning · Facts · Images and charts · Long documents · Following instructions
Mistral Large 3 is the stronger all-rounder.Ministral 3 14B is cheaper.
Scores updated · 26 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. 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 software03Mistral Large 33 of 3 tests
- FactsGetting facts right instead of making them up03Mistral Large 33 of 3 tests
- ReasoningHard problems that need careful thinking01Mistral Large 31 of 3 tests · 2 ties
- Images and chartsUnderstanding pictures, charts and video01Mistral Large 31 of 1 test
- Long documentsFinding answers in very long texts01Mistral Large 31 of 1 test
- Following instructionsDoing exactly what it is asked01Mistral Large 31 of 1 test
- AgentsCarrying out multi-step tasks on its own00Even0 each · 3 ties
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.
Ministral 3 14B costs 80% 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 Mistral Large 3 pulls ahead
- Code for real scientific research problemsSciCode+12.8points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+11.4points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+11.4points ahead
Where Ministral 3 14B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 26 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingMistral Large 3
- SciCodeMistral Large 3 by 12.823.836.6+12.8
- Terminal-Bench HardMistral Large 3 by 11.44.515.9+11.4
- Terminal-Bench 2.1Mistral Large 3 by 2.39.712+2.3
AgentsEven
- τ-Bench V3 · Bankingtie6.65.8tie
- GDPValtie00tie
- Terminal-Bench 4.0tie00tie
ReasoningMistral Large 3
- GPQA DiamondMistral Large 3 by 10.857.268+10.8
- Humanity's Last Examtie4.64.2tie
- CritPttie00tie
FactsMistral Large 3
- AA-Omniscience · AccuracyMistral Large 3 by 11.413.624.9+11.4
- AA-Omniscience · Non-hallucinationMistral Large 3 by 6.57.514+6.5
- Vectara HHEM hallucination ratelower is betterMistral Large 3 by 4.919.414.5+4.9
Images and chartsMistral Large 3
- MMMU-ProMistral Large 3 by 5.949.855.7+5.9
Following instructionsMistral Large 3
- IFBenchMistral Large 3 by 4.23236.2+4.2
Other results11 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.
- MM MTBenchMistral Large 3 by 76.48.584.9+76.4
- LiveCodeBenchMinistral 3 14B by 30.264.634.4+30.2
- AA-OmniscienceMistral Large 3 by 26.8-66.4-39.6+26.8
- Artificial Analysis Coding IndexMistral Large 3 by 5.714.420.1+5.7
- vectara_factual_consistencyMistral Large 3 by 4.980.685.5+4.9
- AA IntelligenceMistral Large 3 by 3.369.3+3.3
- τ²-Bench Telecom (AA run)Ministral 3 14B by 2.627.224.6+2.6
- vectara_avg_summary_lengthMinistral 3 14B by 23.1 rating points135.8112.7+23.1 rating
- vectara_answer_ratetie99.698.8tie
- Arena Hardtie55.155.1tie
- WildBenchtie68.568.5tie
Questions people ask
Which is better, Ministral 3 14B or Mistral Large 3?
Mistral Large 3 wins six of the seven areas where both have results: coding, reasoning, facts, images and charts, long documents and following instructions. Ministral 3 14B wins none, but costs 80% less. They are level on agents.
Which is better for coding?
Mistral Large 3. It wins 3 of the 3 coding tests both models report; Ministral 3 14B wins none.
Which is cheaper?
Ministral 3 14B costs $0.20 per million input tokens and $0.20 per million output tokens; Mistral Large 3 costs $0.50 and $1.50. That makes Ministral 3 14B about 80% cheaper for the same work.
How do you compare the two?
We use the 26 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 11 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.