LFM2.5 24B A2B vs Muse Spark 1.3 Contributor, head to head

Side A

Side B

Overall

Muse Spark 1.3 Contributor is clearly better. It wins 103 of the 105 test cases where they differ.

Learn to build evals like this in the AI Engineering Academy

Task by task

A score turns orange only when that model wins clearly more of the task's test cases. Under each score, the average time per test case on the machine it ran on. Both models spend about the same on thinking per test case.

Thinking tokensAnswer tokens

TaskALFM2.5 24B A2BBMuse Spark 1.3 ContributorLFM2.5 24B A2BMuse Spark 1.3 Contributor

Overall

71m 53s

9.2K8.7K

661m 42s

Agents

Runs multi-step jobs through tools, within a written policy

123m 49s

18.5K14.7K

573m 16s

Structured Extraction

Turns logs, diffs and records into exact JSON

1054.9s

4.4K4.2K

9037.9s

Coding

Writes code and SQL that pass hidden tests

11m 15s

6.1K10.8K

482m 1s

Grounded Answers

Answers from the given documents with sources, never guesses

557.8s

4.7K6.1K

811m 1s

Instruction Following

Applies every rule, including ones changed mid-conversation

162m 44s

13.4K8.7K

841m 18s

Business Logic

Computes billing, scheduling and pricing rules exactly

51m 2s

5K6K

761m 7s

Security & permissions

Ignores planted instructions, still does the allowed work

101m 52s

9K8.2K

761m 59s

Diagnosis & investigation

Chooses diagnostic checks, repairs supported causes and verifies recovery

04m 12s

19.9K17.2K

173m 34s

What you'd run

The build behind each score. Orange marks the faster, smaller, cheaper or longer-context side.

SpecALFM2.5 24B A2BBMuse Spark 1.3 Contributor

Speed

82 tok/s

107 tok/s

Machine

Apple M5 Pro, 48 GB

Hosted

Memory at 32K

25.0 GB

—

Context

128K

Cost per run

Not measured

$0.41

Cost per case

Not measured

$0.0027

Time per case

1m 53s

1m 42s

Quantization

Q8_0

—

Harness

llama.cpp

openrouter

The race

No lead changes

Each model's running score as the test cases go by. The dashed line is a perfect run.

LFM2.5 24B A2BMuse Spark 1.3 Contributor

Every test case

153 test cases

Each square is one test case, colored by the model that scored higher.

Agents112

Structured Extraction017

Coding110

Grounded Answers016

Instruction Following017

Business Logic015

Security & permissions015

Diagnosis & investigation01

2 won by LFM2.5 24B A2B103 won by Muse Spark 1.3 Contributor48 tied