Clover 35B A3B Preview vs DeepSeek V4.1 Flash, head to head

Side A

Side B

Overall

DeepSeek V4.1 Flash is clearly better. It wins 66 of the 70 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. The time under each score is the average per test case, on the machine it ran on. Clover 35B A3B Preview spends 1.5× as many thinking tokens per test case.

Thinking tokensAnswer tokens

TaskAClover 35B A3B PreviewBDeepSeek V4.1 FlashClover 35B A3B PreviewDeepSeek V4.1 Flash

Overall

3035.1s

8.4K6K

643m 35s

Agents

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

2036.8s

9.6K6.1K

464m 42s

Structured Extraction

Turns logs, diffs and records into exact JSON

4327.5s

7.2K4.4K

862m 8s

Coding

Writes code and SQL that pass hidden tests

101m 5s

12.2K9.7K

514m 55s

Grounded Answers

Answers from the given documents with sources, never guesses

4821.0s

5.4K3.9K

762m 17s

Instruction Following

Applies every rule, including ones changed mid-conversation

5332.8s

9.2K7.5K

894m 24s

Business Logic

Computes billing, scheduling and pricing rules exactly

1936.1s

9.8K7K

814m 1s

Security & permissions

Ignores planted instructions, still does the allowed work

4725.6s

5.8K4.4K

693m 3s

Diagnosis & investigation

Chooses diagnostic checks, repairs supported causes and verifies recovery

037.0s

7K3.2K

172m 8s

What you'd run

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

SpecAClover 35B A3B PreviewBDeepSeek V4.1 Flash

Speed

252 tok/s

30 tok/s

Machine

Hosted

Memory at 32K

—

Context

—

Cost per run

Not measured

$0.39

Cost per case

Not measured

$0.0026

Time per case

35.1s

3m 35s

Quantization

—

FP8

Harness

custom

openrouter

The race

No lead changes

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

Clover 35B A3B PreviewDeepSeek V4.1 Flash

Every test case

153 test cases

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

Agents29

Structured Extraction110

Coding09

Grounded Answers17

Instruction Following011

Business Logic013

Security & permissions06

Diagnosis & investigation01

4 won by Clover 35B A3B Preview66 won by DeepSeek V4.1 Flash83 tied