[mle]/DaBench

Apodex 1.1 Mini (free) vs Qwen3.8 Flash

Side A
Model page
Side B
Model page
Overall
25
vs
Overall
32

Qwen3.8 Flash is clearly better. It wins 41 of the 67 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. Apodex 1.1 Mini (free) spends 2.0× as many thinking tokens per test case.

Thinking tokensAnswer tokens
TaskAB
Overall
2535.1s
8.5K, 75% thinking4K, 81% thinking
3258.5s
Agents
2639.4s
6.2K, 69% thinking4.2K, 76% thinking
491m 25s
Structured Extraction
5227.2s
7.7K, 68% thinking3K, 90% thinking
3336.9s
Coding
446.5s
12.1K, 94% thinking5.8K, 61% thinking
471m 21s
Grounded Answers
2926.0s
6.9K, 70% thinking3.1K, 90% thinking
2641.4s
Instruction Following
3644.5s
11.4K, 79% thinking5K, 89% thinking
371m 4s
Business Logic
1435.4s
10K, 56% thinking3.9K, 91% thinking
1746.8s
Security & permissions
4027.3s
5.8K, 76% thinking3.2K, 83% thinking
5054.1s
Diagnosis & investigation
034.6s
5.1K, 84% 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
278 tok/s (better)
77 tok/s
Machine
Hosted
Hosted
Memory at 32K
—
—
Context
256K
1M (better)
Cost per run
Not measured
$0.46
Cost per case
Not measured
$0.003
Time per case
35.1s (better)
58.5s
Quantization
BF16
—
Harness
openrouter
openrouter

The race

6 lead changes

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

Apodex 1.1 Mini (free)Qwen3.8 Flash
025507510050100150Perfect run2532

Every test case

153 test cases

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

26 won by Apodex 1.1 Mini (free)41 won by Qwen3.8 Flash86 tied