Claude Haiku 5.5 vs Qwen3.8 27B, head to head

Side A

Side B

Overall

Claude Haiku 5.5 is clearly better. It wins 64 of the 73 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 27B spends 1.8× as many thinking tokens per test case.

Thinking tokensAnswer tokens

TaskAClaude Haiku 5.5BQwen3.8 27BClaude Haiku 5.5Qwen3.8 27B

Overall

7025.5s

3.9K5.9K

4833.3s

Agents

Runs multi-step jobs through tools, within a written policy

6139.5s

3.7K5.2K

5736.6s

Structured Extraction

Turns logs, diffs and records into exact JSON

9013.5s

2.6K4.6K

7320.6s

Coding

Writes code and SQL that pass hidden tests

8630.6s

6.8K9.3K

3753.0s

Grounded Answers

Answers from the given documents with sources, never guesses

7613.9s

2.2K4.2K

6131.6s

Instruction Following

Applies every rule, including ones changed mid-conversation

9431.5s

4.3K6.9K

6733.6s

Business Logic

Computes billing, scheduling and pricing rules exactly

6726.0s

5.5K7.6K

3632.9s

Security & permissions

Ignores planted instructions, still does the allowed work

6723.4s

2.7K3.7K

5122.6s

Diagnosis & investigation

Chooses diagnostic checks, repairs supported causes and verifies recovery

1725.5s

2.3K5.9K

040.0s

What you'd run

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

SpecAClaude Haiku 5.5BQwen3.8 27B

Speed

225 tok/s

196 tok/s

Machine

Hosted

Memory at 32K

—

Context

256K

Cost per run

$0.42

$1.28

Cost per case

$0.0028

$0.0083

Time per case

25.5s

33.3s

Quantization

—

FP8

Harness

openrouter

The race

2 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 27B

Every test case

153 test cases

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

Agents21

Structured Extraction112

Coding130

Grounded Answers61

Instruction Following120

Business Logic122

Security & permissions73

Diagnosis & investigation10

64 won by Claude Haiku 5.59 won by Qwen3.8 27B80 tied