MiMo-V2.6-Flash vs Mistral Large 4, head to head

Side A

Side B

Overall

MiMo-V2.6-Flash is clearly better. It wins 47 of the 66 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

TaskAMiMo-V2.6-FlashBMistral Large 4MiMo-V2.6-FlashMistral Large 4

Overall

451m 21s

4.2K6.2K

281m 56s

Agents

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

381m 19s

2.2K795

1126.1s

Structured Extraction

Turns logs, diffs and records into exact JSON

6238.0s

2.6K288

153.78s

Coding

Writes code and SQL that pass hidden tests

242m 40s

9.9K9.3K

332m 48s

Grounded Answers

Answers from the given documents with sources, never guesses

6242.1s

2.6K8.1K

382m 45s

Instruction Following

Applies every rule, including ones changed mid-conversation

641m 2s

4.8K12.9K

693m 47s

Business Logic

Computes billing, scheduling and pricing rules exactly

4356.4s

4.4K8.8K

242m 40s

Security & permissions

Ignores planted instructions, still does the allowed work

662m 6s

3.8K3.7K

331m 9s

Diagnosis & investigation

Chooses diagnostic checks, repairs supported causes and verifies recovery

01m 33s

1.5K4.4K

01m 19s

What you'd run

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

SpecAMiMo-V2.6-FlashBMistral Large 4

Speed

66 tok/s

56 tok/s

Machine

Hosted

Memory at 32K

—

Context

1.1M

512K

Cost per run

$0.33

$2.67

Cost per case

$0.0021

$0.017

Time per case

1m 21s

1m 56s

Quantization

BF16

—

Harness

openrouter

The race

No lead changes

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

MiMo-V2.6-FlashMistral Large 4

Every test case

153 test cases

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

Agents71

Structured Extraction122

Coding57

Grounded Answers72

Instruction Following34

Business Logic62

Security & permissions71

Diagnosis & investigation00

47 won by MiMo-V2.6-Flash19 won by Mistral Large 487 tied