MiMo-V2.6-Flash

XiaomiOpen WeightMIT · Uso Comercial

Descripción

MiMo-V2.6-Flash is Xiaomi's MIT-licensed omnimodal sparse Mixture-of-Experts model with 309B total parameters and 15B active parameters. It supports text, image, video, and audio understanding with a 1,048,576-token context window and text output. MiMo-V2.6 scales reinforcement learning across coding, general agents, visual tasks, and cybersecurity. Xiaomi's launch article is dated September 22, 2026; released weights are named MiMo-V2.6-Flash-RL. Benchmark results are Xiaomi-reported release evaluations, distinct from intermediate live-training checkpoints.

Fecha de lanzamiento

2026-09-21

Parámetros

309.0B

Longitud del contexto

1.0M

Modalidades

audio, image, text, video

Radar de capacidades

general

coding

reasoning

scienceest.

agents

multimodal

Science se estima a partir de las puntuaciones científicas de LLM Stats o del razonamiento cuando no hay benchmarks científicos dedicados.

Rankings

Dominio	#Posición	Puntuación	Fuente
Capacidad agéntica	142	26.0	LS
Ranking general	35	40.0	AA
Ranking multimodal	55	56.0	LS
Ciencia	37	38.0	AA

Puntuaciones de benchmarks (LLM Stats)

(LLM Stats (zeroeval))

Agents

Toolathlon-Verified

73.6%Aut.

JobBench

61.2%Aut.

AutomationBench v1.0.6

52.3%Aut.

SEC-bench Pro

47.5%Aut.

Agents' Last Exam

27.6%Aut.

Program Bench

26.0%Aut.

ExploitGym

6.0%Aut.

Code

CyberGym

95.1%Aut.

MiMo Cyber Bench

77.2%Aut.

MiMo Visual Coding

71.5%Aut.

DeepSWE 1.1

67.9%Aut.

MiMo Coding Bench

61.2%Aut.

ExploitBench

25.3%Aut.

Multimodal

OSWorld-Verified

80.8%Aut.

Reasoning

Terminal-Bench 2.1

87.6%Aut.

Terminal-Bench 4.0

28.8%Aut.

Índices de evaluación AA

(Artificial Analysis)

Strategy And Ops

45.7

Finance And Accounting

36.9

Healthcare And Medical

35.1

Legal

33.8

Economics

32.2

Engineering(Artificial Analysis)

31.0

Intelligence Index(Artificial Analysis)

26.7

Puntuaciones por categoría LLM Stats

(LLM Stats (zeroeval))

Multimodal

Vision

Knowledge

Productivity

Reasoning

General

Agents

Code

Tool Calling

Safety

Precios

Precio de entrada$0.14 / 1M tokens

Precio de salida$0.28 / 1M tokens

Precio mixto (3:1)$0.175 / 1M tokens

Precio de lectura caché$0.0028 / 1M tokens

Velocidad

Tokens/seg34.8

Retraso del primer token3.28s

Tiempo hasta la respuesta60.78s

Ranking de Precios por Proveedor

16 proveedores

Más barato: Kilo GatewayMás caro: above.dev

ProveedorEntradaSalida

1Kilo GatewayMás barato

$0.1

$0.28

2XiaomiPRINCIPAL

$0.14

$0.28

3NanoGPT

$0.14

$0.28

4OpenRouter

$0.14

$0.28

5OpenCode Go

$0.14

$0.28

6Ollama Cloud

$0.14

$0.28

7Requesty

$0.14

$0.28

8Vercel AI Gateway

$0.14

$0.28

9DevPass (LLM Gateway)

$0.14

$0.28

10EmpirioLabs AI

$0.14

$0.28

11LLM Gateway

$0.14

$0.28

12Standard Compute

$0.14

$0.28

13Boundless

$0.14

$0.28

14AIHubMix

$0.155

$0.31

15CrossModel

$0.16

$0.32

16above.dev

$0.1692

$0.3385

Comparar precios entre diferentes proveedores de API para este modelo.

Fuentes externas

LLM Stats Artificial Analysis