DeepSeek V4.1 Flash vs DeepSeek V4.1 Flash 1017, head to head

Side A

Side B

Overall

DeepSeek V4.1 Flash is clearly better. It wins 21 of the 31 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

TaskADeepSeek V4.1 FlashBDeepSeek V4.1 Flash 1017DeepSeek V4.1 FlashDeepSeek V4.1 Flash 1017

Overall

643m 35s

6K5.3K

5626.0s

Agents

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

464m 42s

6.1K3.3K

5826.7s

Structured Extraction

Turns logs, diffs and records into exact JSON

862m 8s

4.4K4.2K

8117.2s

Coding

Writes code and SQL that pass hidden tests

514m 55s

9.7K9.4K

4840.8s

Grounded Answers

Answers from the given documents with sources, never guesses

762m 17s

3.9K4.4K

7119.8s

Instruction Following

Applies every rule, including ones changed mid-conversation

894m 24s

7.5K5.5K

7625.6s

Business Logic

Computes billing, scheduling and pricing rules exactly

814m 1s

7K6.6K

6726.6s

Security & permissions

Ignores planted instructions, still does the allowed work

693m 3s

4.4K4.7K

4826.6s

Diagnosis & investigation

Chooses diagnostic checks, repairs supported causes and verifies recovery

172m 8s

3.2K3.2K

020.2s

What you'd run

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

SpecADeepSeek V4.1 FlashBDeepSeek V4.1 Flash 1017

Speed

30 tok/s

243 tok/s

Machine

Hosted

Memory at 32K

—

Context

Cost per run

$0.39

$0.75

Cost per case

$0.0026

$0.0049

Time per case

3m 35s

26.0s

Quantization

FP8

—

Harness

openrouter

The race

4 lead changes

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

DeepSeek V4.1 FlashDeepSeek V4.1 Flash 1017

Every test case

153 test cases

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

Agents25

Structured Extraction21

Coding22

Grounded Answers10

Instruction Following40

Business Logic41

Security & permissions51

Diagnosis & investigation10

21 won by DeepSeek V4.1 Flash10 won by DeepSeek V4.1 Flash 1017122 tied