Granite 4.0 350M vs LFM2.5-8B-A1B
Wins 0 of 6 areas
—
Wins 4 of 6 areas
Coding · Reasoning · Facts · Following instructions
LFM2.5-8B-A1B is the stronger all-rounder.
Scores updated · 18 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 software02LFM2.5-8B-A1B2 of 2 tests
- FactsGetting facts right instead of making them up02LFM2.5-8B-A1B2 of 2 tests
- ReasoningHard problems that need careful thinking01LFM2.5-8B-A1B1 of 3 tests · 2 ties
- Following instructionsDoing exactly what it is asked01LFM2.5-8B-A1B1 of 1 test
- AgentsCarrying out multi-step tasks on its own00Even0 each · 1 tie
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
Agents, long documents and following instructions rest on a single test each.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where LFM2.5-8B-A1B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+41.2points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+38points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+25.6points ahead
Where Granite 4.0 350M pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 18 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingLFM2.5-8B-A1B
- SciCodeLFM2.5-8B-A1B by 6.11.77.8+6.1
- Terminal-Bench HardLFM2.5-8B-A1B by 4.504.5+4.5
ReasoningLFM2.5-8B-A1B
- GPQA DiamondLFM2.5-8B-A1B by 25.625.751.3+25.6
- Humanity's Last Examtie6.46.9tie
- CritPttie00tie
FactsLFM2.5-8B-A1B
- AA-Omniscience · Non-hallucinationLFM2.5-8B-A1B by 41.211.953.1+41.2
- AA-Omniscience · AccuracyLFM2.5-8B-A1B by 5.53.89.3+5.5
Following instructionsLFM2.5-8B-A1B
- IFBenchLFM2.5-8B-A1B by 3817.655.6+38
Other results8 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-OmniscienceLFM2.5-8B-A1B by 47.7-80.9-33.2+47.7
- IFEvalLFM2.5-8B-A1B by 38.353.591.8+38.3
- BFCLv4LFM2.5-8B-A1B by 3613.749.7+36
- τ²-Bench (Retail)LFM2.5-8B-A1B by 33.76.139.8+33.7
- BFCL v3LFM2.5-8B-A1B by 25.239.664.8+25.2
- AA IntelligenceLFM2.5-8B-A1B by 2.44.87.2+2.4
- τ²-Bench Telecom (AA run)LFM2.5-8B-A1B by 1.514.616.1+1.5
- AA Agentic Indextie4.95.4tie
Questions people ask
Which is better, Granite 4.0 350M or LFM2.5-8B-A1B?
LFM2.5-8B-A1B wins four of the six areas where both have results: coding, reasoning, facts and following instructions. Granite 4.0 350M wins none. They are level on agents and long documents.
Which is better for coding?
LFM2.5-8B-A1B. It wins 2 of the 2 coding tests both models report; Granite 4.0 350M wins none.
How do you compare the two?
We use the 18 benchmark tests both models have published scores on. The verdict counts the 10 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 8 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.