LFM3-1.7B vs MiniCPM5-2B
Wins 3 of 4 areas
Coding · Agents · Following instructions
Wins 1 of 4 areas
Reasoning
LFM3-1.7B is the stronger all-rounder.MiniCPM5-2B is better at reasoning.
Scores updated · 20 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.
Agents 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 LFM3-1.7B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+10.9points ahead
- Multi-step tasks using many real software toolsMCP Atlas+6.2points ahead
- Recent programming contest problemsLiveCodeBench v6+3points ahead
Where MiniCPM5-2B pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+6.8points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+1.7points 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.
CodingLFM3-1.7B
- LiveCodeBench v6LFM3-1.7B by 373.570.5+3
- SciCodeLFM3-1.7B by 2.228.526.3+2.2
ReasoningMiniCPM5-2B
- GPQA DiamondMiniCPM5-2B by 6.863.470.2+6.8
- Humanity's Last ExamMiniCPM5-2B by 1.77.28.9+1.7
Following instructionsLFM3-1.7B
- IFBenchLFM3-1.7B by 10.97362.1+10.9
Other results14 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.
- AutomationBench PublicLFM3-1.7B by 36.641.34.7+36.6
- WebArena-VerifiedLFM3-1.7B by 29.338.49.1+29.3
- RULER 128KLFM3-1.7B by 27.665.638+27.6
- BrowseComp-Plus (OpenClaw subset)LFM3-1.7B by 20.23716.8+20.2
- Claw-Eval General (EN), pass^3LFM3-1.7B by 16.851.434.6+16.8
- EnterpriseOps-GymLFM3-1.7B by 14.828.213.4+14.8
- IFEvalLFM3-1.7B by 8.996.787.8+8.9
- BFCLv4LFM3-1.7B by 8.771.462.6+8.7
- Claw-Eval General (EN), averageLFM3-1.7B by 7.968.660.7+7.9
- SWE-bench Multilingual (OpenCode)LFM3-1.7B by 7.32315.7+7.3
- Terminal-Bench 2.1 (OpenCode)LFM3-1.7B by 7.312.95.6+7.3
- HLE (with tools)LFM3-1.7B by 2.413.911.5+2.4
- LiveCodeBench Pro 25Q2LFM3-1.7B by 2.431.729.3+2.4
- τ³-bench (4 domains), averageMiniCPM5-2B by 1.46869.4+1.4
Questions people ask
Which is better, LFM3-1.7B or MiniCPM5-2B?
LFM3-1.7B wins three of the four areas where both have results: coding, agents and following instructions. MiniCPM5-2B wins reasoning.
Which is better for coding?
LFM3-1.7B. It wins 2 of the 2 coding tests both models report; MiniCPM5-2B wins none.
How do you compare the two?
We use the 20 benchmark tests both models have published scores on. The verdict counts the 6 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 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.