[mle]/DaBench

Muse Spark 1.3 Contributor vs Qwen3.8 Flash

Side A
Model page
Side B
Model page
Overall
66
vs
Overall
32

Muse Spark 1.3 Contributor is clearly better. It wins 80 of the 91 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. Muse Spark 1.3 Contributor spends 2.5× as many thinking tokens per test case.

Thinking tokensAnswer tokens
TaskAB
Overall
661m 42s
8.7K, 91% thinking4K, 81% thinking
3258.5s
Agents
573m 16s
14.7K, 88% thinking4.2K, 76% thinking
491m 25s
Structured Extraction
9037.9s
4.2K, 91% thinking3K, 90% thinking
3336.9s
Coding
482m 1s
10.8K, 95% thinking5.8K, 61% thinking
471m 21s
Grounded Answers
811m 1s
6.1K, 91% thinking3.1K, 90% thinking
2641.4s
Instruction Following
841m 18s
8.7K, 91% thinking5K, 89% thinking
371m 4s
Business Logic
761m 7s
6K, 95% thinking3.9K, 91% thinking
1746.8s
Security & permissions
761m 59s
8.2K, 87% thinking3.2K, 83% thinking
5054.1s
Diagnosis & investigation
173m 34s
17.2K, 90% thinking3.3K, 73% thinking
058.4s

What you'd run

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

SpecAB
Speed
107 tok/s (better)
77 tok/s
Machine
Hosted
Hosted
Memory at 32K
—
—
Context
1M
1M
Cost per run
$0.41 (better)
$0.46
Cost per case
$0.0027 (better)
$0.003
Time per case
1m 42s
58.5s (better)
Quantization
—
—
Harness
openrouter
openrouter

The race

2 lead changes

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

Muse Spark 1.3 ContributorQwen3.8 Flash
025507510050100150Perfect run6632

Every test case

153 test cases

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

80 won by Muse Spark 1.3 Contributor11 won by Qwen3.8 Flash62 tied