LFM3-1.7B vs Qwen3.5 4B
Wins 2 of 4 areas
Coding · Following instructions
Wins 2 of 4 areas
Agents · Reasoning
The two are evenly matched.LFM3-1.7B is better at coding and following instructions; Qwen3.5 4B at reasoning and agents.
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.
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 LFM3-1.7B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+21points ahead
- Recent programming contest problemsLiveCodeBench v6+17.7points ahead
- Code for real scientific research problemsSciCode+12.4points ahead
Where Qwen3.5 4B pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+13.7points ahead
- Multi-step tasks using many real software toolsMCP Atlas+8.8points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+2.7points 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 17.773.555.8+17.7
- SciCodeLFM3-1.7B by 12.428.516.1+12.4
ReasoningQwen3.5 4B
- GPQA DiamondQwen3.5 4B by 13.763.477.1+13.7
- Humanity's Last ExamQwen3.5 4B by 2.77.29.9+2.7
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 32.641.38.7+32.6
- BFCLv4LFM3-1.7B by 20.871.450.6+20.8
- RULER 128KQwen3.5 4B by 1965.684.6+19
- WebArena-VerifiedLFM3-1.7B by 16.538.421.9+16.5
- BrowseComp-Plus (OpenClaw subset)LFM3-1.7B by 12.33724.7+12.3
- Claw-Eval General (EN), pass^3LFM3-1.7B by 8.451.443+8.4
- LiveCodeBench Pro 25Q2LFM3-1.7B by 7.431.724.3+7.4
- IFEvalLFM3-1.7B by 6.996.789.8+6.9
- Claw-Eval General (EN), averageLFM3-1.7B by 6.668.662+6.6
- HLE (with tools)LFM3-1.7B by 2.313.911.6+2.3
- EnterpriseOps-GymLFM3-1.7B by 2.228.226+2.2
- τ³-bench (4 domains), averageLFM3-1.7B by 26866+2
- Terminal-Bench 2.1 (OpenCode)tie12.913.5tie
- SWE-bench Multilingual (OpenCode)tie2323.3tie
Questions people ask
Which is better, LFM3-1.7B or Qwen3.5 4B?
LFM3-1.7B and Qwen3.5 4B each win two of the four areas where both have results. LFM3-1.7B wins coding and following instructions; Qwen3.5 4B wins agents and reasoning.
Which is better for coding?
LFM3-1.7B. It wins 2 of the 2 coding tests both models report; Qwen3.5 4B 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.