Gemma 3 4B vs Ministral 3 8B
Wins 0 of 7 areas
—
Wins 5 of 7 areas
Coding · Agents · Facts · Images and charts · Long documents
Ministral 3 8B is the stronger all-rounder.Gemma 3 4B is cheaper.
Scores updated · 25 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 software02Ministral 3 8B2 of 2 tests
- FactsGetting facts right instead of making them up12Ministral 3 8B2 of 3 tests
- AgentsCarrying out multi-step tasks on its own01Ministral 3 8B1 of 2 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video01Ministral 3 8B1 of 1 test
- Long documentsFinding answers in very long texts01Ministral 3 8B1 of 1 test
- ReasoningHard problems that need careful thinking11Even1 each · 1 tie
- Following instructionsDoing exactly what it is asked00Even0 each · 1 tie
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.
Gemma 3 4B costs 50% 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 Ministral 3 8B pulls ahead
- Reasons across sets of long documentsAA-LCR+19points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+18points ahead
- Harder college exam questions with imagesMMMU-Pro+16.1points ahead
Where Gemma 3 4B pulls ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+1points ahead
Every test, side by side
All 25 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingMinistral 3 8B
- SciCodeMinistral 3 8B by 13.47.320.7+13.4
- Terminal-Bench HardMinistral 3 8B by 3.70.84.5+3.7
AgentsMinistral 3 8B
- τ-Bench V3 · BankingMinistral 3 8B by 3.30.43.7+3.3
- GDPValtie00tie
ReasoningEven
- GPQA DiamondMinistral 3 8B by 1829.147.1+18
- Humanity's Last ExamGemma 3 4B by 15.34.3+1
- CritPttie00tie
FactsMinistral 3 8B
- Vectara HHEM hallucination ratelower is betterGemma 3 4B by 15.36.421.7+15.3
- AA-Omniscience · AccuracyMinistral 3 8B by 5.37.713+5.3
- AA-Omniscience · Non-hallucinationMinistral 3 8B by 41.85.8+4
Images and chartsMinistral 3 8B
- MMMU-ProMinistral 3 8B by 16.129.946+16.1
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.
- LiveCodeBenchMinistral 3 8B by 4912.661.6+49
- vectara_answer_rateMinistral 3 8B by 31.867.399.1+31.8
- τ²-Bench Telecom (AA run)Ministral 3 8B by 21.6526.6+21.6
- Arena HardMinistral 3 8B by 19.131.850.9+19.1
- MMLUMinistral 3 8B by 17.758.476.1+17.7
- WildBenchMinistral 3 8B by 17.749.166.8+17.7
- vectara_factual_consistencyGemma 3 4B by 15.393.678.3+15.3
- AA-OmniscienceMinistral 3 8B by 14-82.9-69+14
- Artificial Analysis Coding IndexMinistral 3 8B by 72.79.7+7
- vectara_avg_summary_lengthMinistral 3 8B by 62 rating points77.4139.4+62 rating
- MM MTBenchMinistral 3 8B by 2.95.28.1+2.9
- AA Intelligencetie4.85.5tie
Questions people ask
Which is better, Gemma 3 4B or Ministral 3 8B?
Ministral 3 8B wins five of the seven areas where both have results: coding, agents, facts, images and charts and long documents. Gemma 3 4B wins none, but costs 50% less. They are level on reasoning and following instructions.
Which is better for coding?
Ministral 3 8B. It wins 2 of the 2 coding tests both models report; Gemma 3 4B wins none.
Which is cheaper?
Gemma 3 4B costs $0.05 per million input tokens and $0.10 per million output tokens; Ministral 3 8B costs $0.15 and $0.15. That makes Gemma 3 4B about 50% cheaper for the same work.
How do you compare the two?
We use the 25 benchmark tests both models have published scores on. The verdict counts the 13 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.