Ministral 3 8B vs Qwen3 14B
Wins 1 of 6 areas
Long documents
Wins 5 of 6 areas
Coding · Agents · Reasoning · Facts · Following instructions
Qwen3 14B is the stronger all-rounder.Ministral 3 8B is cheaper and better at long documents.
Scores updated · 26 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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.
- FactsGetting facts right instead of making them up03Qwen3 14B3 of 3 tests
- CodingWriting and fixing software01Qwen3 14B1 of 3 tests · 2 ties
- AgentsCarrying out multi-step tasks on its own01Qwen3 14B1 of 3 tests · 2 ties
- ReasoningHard problems that need careful thinking01Qwen3 14B1 of 3 tests · 2 ties
- Following instructionsDoing exactly what it is asked01Qwen3 14B1 of 1 test
- Long documentsFinding answers in very long texts10Ministral 3 8B1 of 1 test
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 8B costs 83% 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 Qwen3 14B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+18.4points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+13.3points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+11.4points ahead
Where Ministral 3 8B 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.
CodingQwen3 14B
- SciCodeQwen3 14B by 1020.730.7+10
- Terminal-Bench 2.1tie4.14.9tie
- Terminal-Bench Hardtie4.53.8tie
AgentsQwen3 14B
- τ-Bench V3 · BankingQwen3 14B by 1.93.75.6+1.9
- GDPValtie00tie
- Terminal-Bench 4.0tie00tie
ReasoningQwen3 14B
- GPQA DiamondQwen3 14B by 13.347.160.4+13.3
- Humanity's Last Examtie4.34.5tie
- CritPttie00tie
FactsQwen3 14B
- AA-Omniscience · Non-hallucinationQwen3 14B by 18.45.824.2+18.4
- Vectara HHEM hallucination ratelower is betterQwen3 14B by 16.321.75.4+16.3
- AA-Omniscience · AccuracyQwen3 14B by 2.31315.3+2.3
Following instructionsQwen3 14B
- IFBenchQwen3 14B by 11.429.140.5+11.4
Other results12 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.
- AA-OmniscienceQwen3 14B by 20-69-49+20
- vectara_factual_consistencyQwen3 14B by 16.378.394.6+16.3
- AIME 2025Ministral 3 8B by 8.378.770.4+8.3
- Arena HardMinistral 3 8B by 8.250.942.7+8.2
- τ²-Bench Telecom (AA run)Qwen3 14B by 7.926.634.5+7.9
- Artificial Analysis Coding IndexQwen3 14B by 4.19.713.8+4.1
- vectara_avg_summary_lengthMinistral 3 8B by 28.3 rating points139.4111.1+28.3 rating
- AA IntelligenceQwen3 14B by 2.75.58.2+2.7
- AIME 2024Ministral 3 8B by 2.38683.7+2.3
- LiveCodeBenchMinistral 3 8B by 2.361.659.3+2.3
- WildBenchMinistral 3 8B by 1.766.865.1+1.7
- vectara_answer_ratetie99.199.9tie
Questions people ask
Which is better, Ministral 3 8B or Qwen3 14B?
Qwen3 14B wins five of the six areas where both have results: coding, agents, reasoning, facts and following instructions. Ministral 3 8B wins long documents, and costs 83% less.
Which is better for coding?
Qwen3 14B. It wins 1 of the 3 coding tests both models report; Ministral 3 8B wins none, and 2 are ties.
Which is cheaper?
Ministral 3 8B costs $0.15 per million input tokens and $0.15 per million output tokens; Qwen3 14B costs $0.35 and $1.40. That makes Ministral 3 8B about 83% 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 14 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 12 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.