Spark X2.6 4B vs Underdog Saluki 27B, head to head

Side A

Side B

Overall

Underdog Saluki 27B is clearly better. It wins 46 of the 53 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 1.2× as many thinking tokens per test case.

Thinking tokensAnswer tokens

TaskASpark X2.6 4BBUnderdog Saluki 27BSpark X2.6 4BUnderdog Saluki 27B

Overall

151m 3s

8.2K6.9K

335m 15s

Agents

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

141m 32s

11.5K11.8K

409m 2s

Structured Extraction

Turns logs, diffs and records into exact JSON

1039.5s

5.3K4.4K

433m 17s

Coding

Writes code and SQL that pass hidden tests

41m 8s

9.1K9.2K

166m 54s

Grounded Answers

Answers from the given documents with sources, never guesses

1050.2s

6.6K4.5K

433m 22s

Instruction Following

Applies every rule, including ones changed mid-conversation

371m 29s

11.9K8K

466m 5s

Business Logic

Computes billing, scheduling and pricing rules exactly

1036.3s

4.9K5.2K

333m 55s

Security & permissions

Ignores planted instructions, still does the allowed work

3154.3s

7K4.9K

433m 47s

Diagnosis & investigation

Chooses diagnostic checks, repairs supported causes and verifies recovery

01m 30s

11.6K8.7K

06m 39s

What you'd run

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

SpecASpark X2.6 4BBUnderdog Saluki 27B

Speed

136 tok/s

23 tok/s

Machine

RTX 5090, 32 GB

Apple M5 Pro, 48 GB

Memory at 32K

9.4 GB

10.2 GB

Context

605K

256K

Cost per run

$1.61 est.

Not measured

Cost per case

$0.011 est.

Not measured

Time per case

1m 3s

5m 15s

Quantization

BF16

IQ2_XS

Harness

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.

Spark X2.6 4BUnderdog Saluki 27B

Every test case

153 test cases

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

Agents17

Structured Extraction29

Coding16

Grounded Answers07

Instruction Following16

Business Logic05

Security & permissions26

Diagnosis & investigation00

7 won by Spark X2.6 4B46 won by Underdog Saluki 27B100 tied