MiMo-V2.6-Pro vs Qwen3.8 Flash, head to head

Side A

Side B

Overall

MiMo-V2.6-Pro is clearly better. It wins 58 of the 81 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. MiMo-V2.6-Pro spends 1.6× as many thinking tokens per test case.

Thinking tokensAnswer tokens

TaskAMiMo-V2.6-ProBQwen3.8 FlashMiMo-V2.6-ProQwen3.8 Flash

Overall

482m 13s

5.9K4K

3258.5s

Agents

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

474m 48s

8.7K4.2K

491m 25s

Structured Extraction

Turns logs, diffs and records into exact JSON

601m 26s

3.4K3K

3336.9s

Coding

Writes code and SQL that pass hidden tests

242m 58s

10.7K5.8K

471m 21s

Grounded Answers

Answers from the given documents with sources, never guesses

621m 9s

3.2K3.1K

2641.4s

Instruction Following

Applies every rule, including ones changed mid-conversation

591m 23s

4.5K5K

371m 4s

Business Logic

Computes billing, scheduling and pricing rules exactly

571m 52s

7.1K3.9K

1746.8s

Security & permissions

Ignores planted instructions, still does the allowed work

711m 41s

3.6K3.2K

5054.1s

Diagnosis & investigation

Chooses diagnostic checks, repairs supported causes and verifies recovery

02m 54s

6.2K3.3K

058.4s

What you'd run

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

SpecAMiMo-V2.6-ProBQwen3.8 Flash

Speed

55 tok/s

77 tok/s

Machine

Hosted

Memory at 32K

—

Context

1.1M

Cost per run

$1.80

$0.46

Cost per case

$0.012

$0.003

Time per case

2m 13s

58.5s

Quantization

BF16

—

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.

MiMo-V2.6-ProQwen3.8 Flash

Every test case

153 test cases

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

Agents55

Structured Extraction123

Coding29

Grounded Answers111

Instruction Following92

Business Logic121

Security & permissions72

Diagnosis & investigation00

58 won by MiMo-V2.6-Pro23 won by Qwen3.8 Flash72 tied