Granite 4.0 H Tiny vs LFM2-2.6B
Wins 1 of 3 areas
Coding
Wins 2 of 3 areas
Reasoning · Following instructions
LFM2-2.6B is the stronger all-rounder.Granite 4.0 H Tiny is better at coding.
Scores updated · 13 tests both models report · How we compare
Where each one wins
Tests won in each of the three 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.
Coding, reasoning 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 LFM2-2.6B pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+4.1points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+1.1points ahead
Where Granite 4.0 H Tiny pulls ahead
- Recent programming contest problemsLiveCodeBench v6+12.3points ahead
Every test, side by side
All 13 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Following instructionsLFM2-2.6B
- IFBenchLFM2-2.6B by 1.118.419.5+1.1
Other results10 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.
- HumanEval+Granite 4.0 H Tiny by 15.973.857.9+15.9
- LCB v5Granite 4.0 H Tiny by 12.927.314.4+12.9
- MMLU-ProGranite 4.0 H Tiny by 63226+6
- MATH-500 (EM)LFM2-2.6B by 5.458.263.6+5.4
- MMLUGranite 4.0 H Tiny by 2.466.864.4+2.4
- GSMPlusLFM2-2.6B by 1.759.160.8+1.7
- IFEvalGranite 4.0 H Tiny by 1.581.179.6+1.5
- MMMLUtie56.155.4tie
- MGSMtie73.774.3tie
- GSM8Ktie82.682.4tie
Questions people ask
Which is better, Granite 4.0 H Tiny or LFM2-2.6B?
LFM2-2.6B wins two of the three areas where both have results: reasoning and following instructions. Granite 4.0 H Tiny wins coding.
Which is better for coding?
Granite 4.0 H Tiny. It wins the one coding test both models report.
How do you compare the two?
We use the 13 benchmark tests both models have published scores on. The verdict counts the 3 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 10 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.