NOUMENOpen app
NOUMEN

xiaomi · Open weights

MiMo V2.6 Flash Abliterated

Added Sep 25, 2026Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

A MiMo V2.6 Flash finetune with refusal-direction ablation, image input, a 1M-token context window, separate reasoning output, and tool calling.

  • Reasoning
  • Tools
  • Vision
Input
Text, Images
Output
Text
Reasoning effort
none · high
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.50
Output
$1.50
Cache read
$0.28

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
65.5K
Average output (7 days)
1.3K tokens
Longer than 80% of 465 models

Side by side

MiMo V2.6 Flash Abliterated compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

MiMo V2.6 Flash AbliteratedxiaomiThis modelGLM 4.6 OriginalZ.aiRemove
ProviderxiaomiZ.ai
Discovery tierOpen weightsOpen weights
AddedSep 25, 2026Dec 11, 2025
Limits
Context window1.05MBest256K
Max output65.54KBest65.54K
Average output1.3K tokens1.3K tokens
Pricing, USD per 1M tokens
Input$0.50$0.50
Output$1.50Best$2.00
Cache read$0.28$0.10Best
Performance
Throughput130.9 tok/s—
Time to first token2.4 s—
Request success (24h)98.4%99.4%Best
Benchmark indexes
Intelligence Index—14.9
Math Index—44.3
Capabilities
ReasoningYesYes
ToolsYesYes
VisionYesNo
Reasoning effortnone, high—

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • MiMo V2.6 Flash Abliterated
  • GLM 4.6 Original

Intelligence Index

  • MiMo V2.6 Flash Abliterated—
  • GLM 4.6 Original14.9
3.8Catalog models by score55.1

Indexes

  • Math Index

    Composite of the math evaluations

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 44.3

Document reasoning

  • AA-LCRv1.1

    Reasoning across long documents

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 26.3%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 5.5%

  • GPQA Diamond

    Graduate-level science questions

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 63.2%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 78.4%

  • IFBench

    Following precise output instructions

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 36.7%

Coding

  • LiveCodeBench

    Recent competitive programming problems

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 56.1%

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 44.3%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 28.8%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 76.9%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • MiMo V2.6 Flash Abliterated: —
    • GLM 4.6 Original: 26.3%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
130.9 tok/s
Faster than 72% of 250 models
Time to first token
2.4 s
Quicker than 70% of 250 models
Prompt cache hits
95.5%
About $0.29 per 1M input tokens with caching
Request success (24h)
98.4%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 0.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "xiaomi/mimo-v2.6-flash-abliterated",
    "reasoning_effort": "none",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="xiaomi/mimo-v2.6-flash-abliterated",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

xiaomi

Other xiaomi models, starting with the same family.