LFM2.5-1.2B-Thinking vs LFM2.5-1.2B-Instruct
Wins 2 of 6 areas
Coding · Reasoning
Wins 2 of 6 areas
Facts · Following instructions
The two are evenly matched.LFM2.5-1.2B-Thinking is better at reasoning and coding; LFM2.5-1.2B-Instruct at facts and following instructions.
Scores updated · 20 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.
- ReasoningHard problems that need careful thinking10LFM2.5-1.2B-Thinking1 of 3 tests · 2 ties
- CodingWriting and fixing software10LFM2.5-1.2B-Thinking1 of 2 tests · 1 tie
- FactsGetting facts right instead of making them up01LFM2.5-1.2B-Instruct1 of 2 tests · 1 tie
- Following instructionsDoing exactly what it is asked01LFM2.5-1.2B-Instruct1 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 LFM2.5-1.2B-Thinking pulls ahead
- Code for real scientific research problemsSciCode+1.9points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+1.3points ahead
Where LFM2.5-1.2B-Instruct pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+9.3points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+2points ahead
Every test, side by side
All 20 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingLFM2.5-1.2B-Thinking
- SciCodeLFM2.5-1.2B-Thinking by 1.94.22.3+1.9
- Terminal-Bench Hardtie00tie
ReasoningLFM2.5-1.2B-Thinking
- GPQA DiamondLFM2.5-1.2B-Thinking by 1.333.932.6+1.3
- Humanity's Last Examtie6.26.7tie
- CritPttie00tie
FactsLFM2.5-1.2B-Instruct
- AA-Omniscience · Non-hallucinationLFM2.5-1.2B-Instruct by 9.35.614.9+9.3
- AA-Omniscience · Accuracytie7.47tie
Following instructionsLFM2.5-1.2B-Instruct
- IFBenchLFM2.5-1.2B-Instruct by 241.843.8+2
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.
- MATH-500 (EM)LFM2.5-1.2B-Thinking by 82.2885.8+82.2
- GSM8KLFM2.5-1.2B-Thinking by 81.385.64.3+81.3
- AIME25 no toolsLFM2.5-1.2B-Thinking by 17.731.714+17.7
- τ²-Bench Telecom (AA run)LFM2.5-1.2B-Thinking by 8.819.610.8+8.8
- BFCL v3LFM2.5-1.2B-Thinking by 7.95749.1+7.9
- AA-OmniscienceLFM2.5-1.2B-Instruct by 7.8-80-72.1+7.8
- MMLU-ProLFM2.5-1.2B-Thinking by 5.349.644.4+5.3
- AA Agentic IndexLFM2.5-1.2B-Thinking by 2.96.53.6+2.9
- IFEvalLFM2.5-1.2B-Thinking by 2.288.486.2+2.2
- AA Intelligencetie5.25.2tie
Questions people ask
Which is better, LFM2.5-1.2B-Thinking or LFM2.5-1.2B-Instruct?
LFM2.5-1.2B-Thinking and LFM2.5-1.2B-Instruct each win two of the six areas where both have results. LFM2.5-1.2B-Thinking wins coding and reasoning; LFM2.5-1.2B-Instruct wins facts and following instructions. They are level on agents and long documents.
Which is better for coding?
LFM2.5-1.2B-Thinking. It wins 1 of the 2 coding tests both models report; LFM2.5-1.2B-Instruct wins none, and 1 is a tie.
How do you compare the two?
We use the 20 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 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.