LFM2-24B-A2B vs Qwen2.5-VL-32B

Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.

LFM2-24B-A2B and Qwen2.5-VL-32B finish on the same overall score, so this is less about a single winner and more about where the edge shows up. The headline says tie; the benchmark table is where the real choice happens.

Quick Verdict

Treat this as a split decision. LFM2-24B-A2B makes more sense if coding is the priority; Qwen2.5-VL-32B is the better fit if multimodal & grounded is the priority.

Agentic

Qwen2.5-VL-32B

LFM2-24B-A2B

33.4

Qwen2.5-VL-32B

33.5

30
Terminal-Bench 2.0
30
38
BrowseComp
41
34
OSWorld-Verified
32

Coding

LFM2-24B-A2B

LFM2-24B-A2B

18

Qwen2.5-VL-32B

14.3

42
HumanEval
35
18
SWE-bench Verified
17
17
LiveCodeBench
11
19
SWE-bench Pro
17

Multimodal & Grounded

Qwen2.5-VL-32B

LFM2-24B-A2B

41.7

Qwen2.5-VL-32B

52.2

39
MMMU-Pro
58
45
OfficeQA Pro
45

Reasoning

LFM2-24B-A2B

LFM2-24B-A2B

46.6

Qwen2.5-VL-32B

43.2

44
SimpleQA
41
42
MuSR
39
63
BBH
61
48
LongBench v2
42
45
MRCRv2
43

Knowledge

LFM2-24B-A2B

LFM2-24B-A2B

35.6

Qwen2.5-VL-32B

34.7

46
MMLU
43
45
GPQA
42
43
SuperGPQA
40
41
OpenBookQA
38
51
MMLU-Pro
53
4
HLE
2
43
FrontierScience
43

Instruction Following

LFM2-24B-A2B

LFM2-24B-A2B

68

Qwen2.5-VL-32B

67

68
IFEval
67

Multilingual

LFM2-24B-A2B

LFM2-24B-A2B

61.4

Qwen2.5-VL-32B

60.4

64
MGSM
63
60
MMLU-ProX
59

Mathematics

LFM2-24B-A2B

LFM2-24B-A2B

50.4

Qwen2.5-VL-32B

49.7

46
AIME 2023
43
48
AIME 2024
45
47
AIME 2025
44
42
HMMT Feb 2023
39
44
HMMT Feb 2024
41
43
HMMT Feb 2025
40
45
BRUMO 2025
42
57
MATH-500
59

Frequently Asked Questions

Which is better, LFM2-24B-A2B or Qwen2.5-VL-32B?

LFM2-24B-A2B and Qwen2.5-VL-32B are tied on overall score, so the right pick depends on which category matters most for your use case.

Which is better for knowledge tasks, LFM2-24B-A2B or Qwen2.5-VL-32B?

LFM2-24B-A2B has the edge for knowledge tasks in this comparison, averaging 35.6 versus 34.7. Inside this category, MMLU is the benchmark that creates the most daylight between them.

Which is better for coding, LFM2-24B-A2B or Qwen2.5-VL-32B?

LFM2-24B-A2B has the edge for coding in this comparison, averaging 18 versus 14.3. Inside this category, HumanEval is the benchmark that creates the most daylight between them.

Which is better for math, LFM2-24B-A2B or Qwen2.5-VL-32B?

LFM2-24B-A2B has the edge for math in this comparison, averaging 50.4 versus 49.7. Inside this category, AIME 2023 is the benchmark that creates the most daylight between them.

Which is better for reasoning, LFM2-24B-A2B or Qwen2.5-VL-32B?

LFM2-24B-A2B has the edge for reasoning in this comparison, averaging 46.6 versus 43.2. Inside this category, LongBench v2 is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, LFM2-24B-A2B or Qwen2.5-VL-32B?

Qwen2.5-VL-32B has the edge for agentic tasks in this comparison, averaging 33.5 versus 33.4. Inside this category, BrowseComp is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, LFM2-24B-A2B or Qwen2.5-VL-32B?

Qwen2.5-VL-32B has the edge for multimodal and grounded tasks in this comparison, averaging 52.2 versus 41.7. Inside this category, MMMU-Pro is the benchmark that creates the most daylight between them.

Which is better for instruction following, LFM2-24B-A2B or Qwen2.5-VL-32B?

LFM2-24B-A2B has the edge for instruction following in this comparison, averaging 68 versus 67. Inside this category, IFEval is the benchmark that creates the most daylight between them.

Which is better for multilingual tasks, LFM2-24B-A2B or Qwen2.5-VL-32B?

LFM2-24B-A2B has the edge for multilingual tasks in this comparison, averaging 61.4 versus 60.4. Inside this category, MGSM is the benchmark that creates the most daylight between them.

Last updated: March 12, 2026

Weekly LLM Benchmark Digest

Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.

Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.