Gemma 4 E4B vs LFM2.5-2.6B
Wins 1 of 6 areas
Long documents
Wins 3 of 6 areas
Coding · Agents · Following instructions
LFM2.5-2.6B wins more areas, narrowly.Gemma 4 E4B is better at long documents.
Scores updated · 24 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.
- CodingWriting and fixing software12LFM2.5-2.6B2 of 3 tests
- AgentsCarrying out multi-step tasks on its own01LFM2.5-2.6B1 of 2 tests · 1 tie
- Following instructionsDoing exactly what it is asked01LFM2.5-2.6B1 of 1 test
- Long documentsFinding answers in very long texts10Gemma 4 E4B1 of 1 test
- ReasoningHard problems that need careful thinking11Even1 each · 1 tie
- FactsGetting facts right instead of making them up11Even1 each
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.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where LFM2.5-2.6B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+15points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+14.9points ahead
- Recent programming contest problemsLiveCodeBench v6+7.4points ahead
Where Gemma 4 E4B pulls ahead
- Reasons across sets of long documentsAA-LCR+26.3points ahead
- Code for real scientific research problemsSciCode+10points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+4.2points ahead
Every test, side by side
All 24 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingLFM2.5-2.6B
- SciCodeGemma 4 E4B by 1024.414.4+10
- LiveCodeBench v6LFM2.5-2.6B by 7.45259.4+7.4
- Terminal-Bench 2.1LFM2.5-2.6B by 2.61.94.5+2.6
ReasoningEven
- Humanity's Last ExamLFM2.5-2.6B by 2.43.86.2+2.4
- GPQA DiamondGemma 4 E4B by 1.857.655.8+1.8
- CritPttie0.60tie
FactsEven
- AA-Omniscience · Non-hallucinationLFM2.5-2.6B by 14.969.184+14.9
- AA-Omniscience · AccuracyGemma 4 E4B by 4.28.64.3+4.2
Following instructionsLFM2.5-2.6B
- IFBenchLFM2.5-2.6B by 1544.259.2+15
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.
- MATH500 (Pass@1)Gemma 4 E4B by 59.6655.4+59.6
- AIME25 no toolsLFM2.5-2.6B by 17.634.351.9+17.6
- BFCLv4LFM2.5-2.6B by 16.94056.9+16.9
- PinchBenchLFM2.5-2.6B by 13.155.168.2+13.1
- BrowseComp+ (OpenClaw)LFM2.5-2.6B by 1115.926.9+11
- AA-OmniscienceLFM2.5-2.6B by 8.8-19.7-10.9+8.8
- Claw-Eval AvgLFM2.5-2.6B by 4.95862.9+4.9
- LiveCodeBenchV6 no toolsGemma 4 E4B by 4.463.859.4+4.4
- Artificial Analysis Coding IndexGemma 4 E4B by 1.79.47.7+1.7
- τ³-Bench BankingLFM2.5-2.6B by 1.64.15.7+1.6
- AA Agentic Indextie1.82.4tie
- AA Intelligencetie8.98.4tie
Questions people ask
Which is better, Gemma 4 E4B or LFM2.5-2.6B?
LFM2.5-2.6B wins three of the six areas where both have results: coding, agents and following instructions. Gemma 4 E4B wins long documents. They are level on reasoning and facts.
Which is better for coding?
LFM2.5-2.6B. It wins 2 of the 3 coding tests both models report; Gemma 4 E4B wins 1.
How do you compare the two?
We use the 24 benchmark tests both models have published scores on. The verdict counts the 12 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.