[mle]/DaBench

Clover 35B A3B Preview vs MiniMax M3.1 Flash

Side A
Model page
Side B
Model page
Overall
30
vs
Overall
35

MiniMax M3.1 Flash is clearly better. It wins 37 of the 61 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. The time under each score is the average per test case, on the machine it ran on. Clover 35B A3B Preview spends 4.7× as many thinking tokens per test case.

Thinking tokensAnswer tokens
TaskAB
Overall
3035.1s
8.4K, 95% thinking2.3K, 73% thinking
3521.3s
Agents
2036.8s
9.6K, 91% thinking3.1K, 71% thinking
3637.0s
Structured Extraction
4327.5s
7.2K, 97% thinking1.1K, 74% thinking
487.96s
Coding
101m 5s
12.2K, 100% thinking4.6K, 66% thinking
5840.6s
Grounded Answers
4821.0s
5.4K, 94% thinking1.2K, 71% thinking
339.75s
Instruction Following
5332.8s
9.2K, 95% thinking1.7K, 70% thinking
5314.7s
Business Logic
1936.1s
9.8K, 99% thinking3.5K, 94% thinking
2425.9s
Security & permissions
4725.6s
5.8K, 92% thinking1.3K, 60% thinking
3012.7s
Diagnosis & investigation
037.0s
7K, 82% thinking2.1K, 51% thinking
022.8s

What you'd run

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

SpecAB
Speed
252 tok/s (better)
135 tok/s
Machine
Hosted
Hosted
Memory at 32K
—
—
Context
—
1M
Cost per run
Not measured
Not measured
Cost per case
Not measured
Not measured
Time per case
35.1s
21.3s (better)
Quantization
—
—
Harness
custom
openrouter

The race

No lead changes

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

Clover 35B A3B PreviewMiniMax M3.1 Flash
0255075100501001503035

Every test case

153 test cases

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

24 won by Clover 35B A3B Preview37 won by MiniMax M3.1 Flash92 tied