[mle]/DaBench

Claude Haiku 5.5 vs Qwen3.8 Flash

Side A
Model page
Side B
Model page
Overall
70
vs
Overall
32

Claude Haiku 5.5 is clearly better. It wins 82 of the 87 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. Both models spend about the same on thinking per test case.

Thinking tokensAnswer tokens
TaskAB
Overall
7025.5s
3.9K, 78% thinking4K, 81% thinking
3258.5s
Agents
6139.5s
3.7K, 61% thinking4.2K, 76% thinking
491m 25s
Structured Extraction
9013.5s
2.6K, 83% thinking3K, 90% thinking
3336.9s
Coding
8630.6s
6.8K, 76% thinking5.8K, 61% thinking
471m 21s
Grounded Answers
7613.9s
2.2K, 78% thinking3.1K, 90% thinking
2641.4s
Instruction Following
9431.5s
4.3K, 82% thinking5K, 89% thinking
371m 4s
Business Logic
6726.0s
5.5K, 95% thinking3.9K, 91% thinking
1746.8s
Security & permissions
6723.4s
2.7K, 70% thinking3.2K, 83% thinking
5054.1s
Diagnosis & investigation
1725.5s
2.3K, 50% 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
225 tok/s (better)
77 tok/s
Machine
Hosted
Hosted
Memory at 32K
—
—
Context
1M
1M
Cost per run
$0.42 (better)
$0.46
Cost per case
$0.0028 (better)
$0.003
Time per case
25.5s (better)
58.5s
Quantization
—
—
Harness
openrouter
openrouter

The race

No lead changes

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

Claude Haiku 5.5Qwen3.8 Flash
025507510050100150Perfect run7032

Every test case

153 test cases

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

82 won by Claude Haiku 5.55 won by Qwen3.8 Flash66 tied