[mle]/DaBench

MiniMax M3.1 Flash vs Muse Spark 1.3 Contributor

Side A
Model page
Side B
Model page
Overall
35
vs
Overall
66

Muse Spark 1.3 Contributor is clearly better. It wins 61 of the 71 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 4.6× as many thinking tokens per test case.

Thinking tokensAnswer tokens
TaskAB
Overall
3521.3s
2.3K, 73% thinking8.7K, 91% thinking
661m 42s
Agents
3637.0s
3.1K, 71% thinking14.7K, 88% thinking
573m 16s
Structured Extraction
487.96s
1.1K, 74% thinking4.2K, 91% thinking
9037.9s
Coding
5840.6s
4.6K, 66% thinking10.8K, 95% thinking
482m 1s
Grounded Answers
339.75s
1.2K, 71% thinking6.1K, 91% thinking
811m 1s
Instruction Following
5314.7s
1.7K, 70% thinking8.7K, 91% thinking
841m 18s
Business Logic
2425.9s
3.5K, 94% thinking6K, 95% thinking
761m 7s
Security & permissions
3012.7s
1.3K, 60% thinking8.2K, 87% thinking
761m 59s
Diagnosis & investigation
022.8s
2.1K, 51% thinking17.2K, 90% thinking
173m 34s

What you'd run

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

SpecAB
Speed
135 tok/s (better)
107 tok/s
Machine
Hosted
Hosted
Memory at 32K
—
—
Context
1M
1M
Cost per run
Not measured
$0.41
Cost per case
Not measured
$0.0027
Time per case
21.3s (better)
1m 42s
Quantization
—
—
Harness
openrouter
openrouter

The race

1 lead change

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

MiniMax M3.1 FlashMuse Spark 1.3 Contributor
025507510050100150Perfect run3566

Every test case

153 test cases

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

10 won by MiniMax M3.1 Flash61 won by Muse Spark 1.3 Contributor82 tied