[mle]/DaBench

Claude Haiku 5.5 vs Spark X2.6 4B

Side A
Model page
Side B
Model page
Overall
70
vs
Overall
15

Claude Haiku 5.5 is clearly better. It wins 99 of the 101 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. Spark X2.6 4B spends 2.3× as many thinking tokens per test case.

Thinking tokensAnswer tokens
TaskAB
Overall
7025.5s
3.9K, 78% thinking8.2K, 84% thinking
151m 3s
Agents
6139.5s
3.7K, 61% thinking11.5K, 97% thinking
141m 32s
Structured Extraction
9013.5s
2.6K, 83% thinking5.3K, 78% thinking
1039.5s
Coding
8630.6s
6.8K, 76% thinking9.1K, 48% thinking
41m 8s
Grounded Answers
7613.9s
2.2K, 78% thinking6.6K, 86% thinking
1050.2s
Instruction Following
9431.5s
4.3K, 82% thinking11.9K, 90% thinking
371m 29s
Business Logic
6726.0s
5.5K, 95% thinking4.9K, 86% thinking
1036.3s
Security & permissions
6723.4s
2.7K, 70% thinking7K, 94% thinking
3154.3s
Diagnosis & investigation
1725.5s
2.3K, 50% thinking11.6K, 98% thinking
01m 30s

What you'd run

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

SpecAB
Speed
225 tok/s
136 tok/s
Machine
Hosted
RTX 5090, 32 GB
Memory at 32K
—
9.4 GB
Context
1M (better)
605K
Cost per run
$0.42 (better)
$1.61 est.
Cost per case
$0.0028 (better)
$0.011 est.
Time per case
25.5s (better)
1m 3s
Quantization
—
BF16
Harness
openrouter
llama.cpp

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 5.5Spark X2.6 4B
0255075100501001507015

Every test case

153 test cases

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

99 won by Claude Haiku 5.52 won by Spark X2.6 4B52 tied