Gemma 3 4B vs LFM2-8B-A1B (Non-Reasoning)
Wins 1 of 4 areas
Following instructions
Wins 3 of 4 areas
Coding · Reasoning · Facts
LFM2-8B-A1B (Non-Reasoning) is the stronger all-rounder.Gemma 3 4B is better at following instructions.
Scores updated · 21 tests both models report · How we compare
Where each one wins
Tests won in each of the four 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 software01LFM2-8B-A1B (Non-Reasoning)1 of 3 tests · 2 ties
- ReasoningHard problems that need careful thinking01LFM2-8B-A1B (Non-Reasoning)1 of 3 tests · 2 ties
- FactsGetting facts right instead of making them up01LFM2-8B-A1B (Non-Reasoning)1 of 2 tests · 1 tie
- Following instructionsDoing exactly what it is asked10Gemma 3 4B1 of 1 test
Following instructions rests on a single test.
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.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where LFM2-8B-A1B (Non-Reasoning) pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+5.9points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+5.3points ahead
- Recent programming contest problemsLiveCodeBench v6+2.1points ahead
Where Gemma 3 4B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+2points ahead
Every test, side by side
All 21 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingLFM2-8B-A1B (Non-Reasoning)
- LiveCodeBench v6LFM2-8B-A1B (Non-Reasoning) by 2.118.921+2.1
- Terminal-Bench Hardtie0.80tie
- SciCodetie7.36.8tie
ReasoningLFM2-8B-A1B (Non-Reasoning)
- GPQA DiamondLFM2-8B-A1B (Non-Reasoning) by 5.329.134.4+5.3
- Humanity's Last Examtie5.34.9tie
- CritPttie00tie
FactsLFM2-8B-A1B (Non-Reasoning)
- AA-Omniscience · Non-hallucinationLFM2-8B-A1B (Non-Reasoning) by 5.91.87.7+5.9
- AA-Omniscience · Accuracytie7.78.2tie
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.
- MGSMGemma 3 4B by 14.987.372.4+14.9
- IFEvalGemma 3 4B by 12.690.277.6+12.6
- HumanEval+LFM2-8B-A1B (Non-Reasoning) by 6.762.869.5+6.7
- AA-OmniscienceLFM2-8B-A1B (Non-Reasoning) by 6.4-82.9-76.5+6.4
- MMLULFM2-8B-A1B (Non-Reasoning) by 6.458.464.8+6.4
- MMLU-ProGemma 3 4B by 6.243.637.4+6.2
- τ²-Bench Telecom (AA run)LFM2-8B-A1B (Non-Reasoning) by 5.5510.5+5.5
- MMMLULFM2-8B-A1B (Non-Reasoning) by 5.250.155.3+5.2
- GSMPlusGemma 3 4B by 3.668.464.8+3.6
- LCB v5LFM2-8B-A1B (Non-Reasoning) by 2.319.121.4+2.3
- MATH-500 (EM)LFM2-8B-A1B (Non-Reasoning) by 173.274.2+1
- AA Intelligencetie4.84.9tie
Questions people ask
Which is better, Gemma 3 4B or LFM2-8B-A1B (Non-Reasoning)?
LFM2-8B-A1B (Non-Reasoning) wins three of the four areas where both have results: coding, reasoning and facts. Gemma 3 4B wins following instructions.
Which is better for coding?
LFM2-8B-A1B (Non-Reasoning). It wins 1 of the 3 coding tests both models report; Gemma 3 4B wins none, and 2 are ties.
How do you compare the two?
We use the 21 benchmark tests both models have published scores on. The verdict counts the 9 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.