Gemma 3 12B vs Gemma 4 26B A4B
Wins 0 of 7 areas
—
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents · Following instructions
Gemma 4 26B A4B is the stronger all-rounder.Gemma 3 12B is cheaper.
Scores updated · 30 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 software03Gemma 4 26B A4B3 of 3 tests
- Images and chartsUnderstanding pictures, charts and video03Gemma 4 26B A4B3 of 3 tests
- ReasoningHard problems that need careful thinking02Gemma 4 26B A4B2 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up02Gemma 4 26B A4B2 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own02Gemma 4 26B A4B2 of 2 tests
- Long documentsFinding answers in very long texts01Gemma 4 26B A4B1 of 1 test
- Following instructionsDoing exactly what it is asked01Gemma 4 26B A4B1 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.
Gemma 3 12B costs 62% 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 Gemma 4 26B A4B pulls ahead
- Reasons across sets of long documentsAA-LCR+57.4points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+44.3points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+35.7points ahead
Where Gemma 3 12B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 30 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemma 4 26B A4B
- Terminal-Bench 2.1Gemma 4 26B A4B by 39039+39
- SciCodeGemma 4 26B A4B by 23.616.440+23.6
- Terminal-Bench HardGemma 4 26B A4B by 12.80.813.6+12.8
AgentsGemma 4 26B A4B
- τ-Bench V3 · BankingGemma 4 26B A4B by 11.20.812+11.2
- GDPValGemma 4 26B A4B by 3.403.4+3.4
ReasoningGemma 4 26B A4B
- GPQA DiamondGemma 4 26B A4B by 44.334.979.2+44.3
- Humanity's Last ExamGemma 4 26B A4B by 15.14.219.3+15.1
- CritPttie00tie
FactsGemma 4 26B A4B
- AA-Omniscience · Non-hallucinationGemma 4 26B A4B by 112.613.6+11
- AA-Omniscience · AccuracyGemma 4 26B A4B by 8.810.319.1+8.8
- Vectara HHEM hallucination ratelower is bettertie4.45.2tie
Images and chartsGemma 4 26B A4B
Following instructionsGemma 4 26B A4B
- IFBenchGemma 4 26B A4B by 35.736.772.4+35.7
Other results14 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.
- MATH-VisionGemma 4 26B A4B by 50.332.182.4+50.3
- BIG-Bench Extra HardGemma 4 26B A4B by 48.516.364.8+48.5
- HumanEvalGemma 3 12B by 35.485.450+35.4
- Artificial Analysis Coding IndexGemma 4 26B A4B by 33.55.839.3+33.5
- τ²-Bench Telecom (AA run)Gemma 4 26B A4B by 32.810.843.6+32.8
- AA-OmniscienceGemma 4 26B A4B by 26.3-77.1-50.8+26.3
- vectara_avg_summary_lengthGemma 3 12B by 22.689.767.1+22.6
- MMLU-ProGemma 4 26B A4B by 2260.682.6+22
- AA IntelligenceGemma 4 26B A4B by 12.93.816.7+12.9
- MBPPGemma 3 12B by 4.67368.4+4.6
- AI2DGemma 4 26B A4B by 4.184.288.3+4.1
- IFEvalGemma 4 26B A4B by 2.588.991.4+2.5
- vectara_answer_rateGemma 4 26B A4B by 2.497.499.8+2.4
- vectara_factual_consistencytie95.694.8tie
Questions people ask
Which is better, Gemma 3 12B or Gemma 4 26B A4B?
Gemma 4 26B A4B wins all seven areas where both have results: coding, agents, reasoning, facts, images and charts, long documents and following instructions. Gemma 3 12B wins none, but costs 62% less.
Which is better for coding?
Gemma 4 26B A4B. It wins 3 of the 3 coding tests both models report; Gemma 3 12B wins none.
Which is cheaper?
Gemma 3 12B costs $0.05 per million input tokens and $0.15 per million output tokens; Gemma 4 26B A4B costs $0.13 and $0.40. That makes Gemma 3 12B about 62% cheaper for the same work.
How do you compare the two?
We use the 30 benchmark tests both models have published scores on. The verdict counts the 16 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 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.