[mle]/DaBench

MiniMax M3.1 Flash vs Qwen3.8 Flash Next

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

Too close to call. MiniMax M3.1 Flash wins 33 of the 62 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. Qwen3.8 Flash Next spends 1.6× as many thinking tokens per test case.

Thinking tokensAnswer tokens
TaskAB
Overall
3521.3s
2.3K, 73% thinking3.3K, 80% thinking
351m 0s
Agents
3637.0s
3.1K, 71% thinking4.6K, 78% thinking
522m 4s
Structured Extraction
487.96s
1.1K, 74% thinking3.1K, 90% thinking
5042.2s
Coding
5840.6s
4.6K, 66% thinking6.2K, 61% thinking
521m 36s
Grounded Answers
339.75s
1.2K, 71% thinking3.1K, 88% thinking
5147.0s
Instruction Following
5314.7s
1.7K, 70% thinking3.3K, 90% thinking
431m 0s
Business Logic
2425.9s
3.5K, 94% thinking3.5K, 91% thinking
2156.0s
Security & permissions
3012.7s
1.3K, 60% thinking315, 92% thinking
1011.9s
Diagnosis & investigation
022.8s
2.1K, 51% thinking0, no thinking
02.15s

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)
68 tok/s
Machine
Hosted
Hosted
Memory at 32K
—
—
Context
1M
1M
Cost per run
Not measured
$0.36
Cost per case
Not measured
$0.0024
Time per case
21.3s (better)
1m 0s
Quantization
—
—
Harness
openrouter
openrouter

The race

4 lead changes

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

MiniMax M3.1 FlashQwen3.8 Flash Next
025507510050100150Perfect run3535

Every test case

153 test cases

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

33 won by MiniMax M3.1 Flash29 won by Qwen3.8 Flash Next91 tied