# Benchmarks sourced from MiMo-V2.6-Flash-RL use its "MiMo-V2.5 Pro" comparison column.
name = "MiMo-V2.5-Pro"
description = "Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution"
family = "mimo"
release_date = "2026-04-22"
last_updated = "2026-04-22"
attachment = false
reasoning = true
temperature = true
tool_call = true
knowledge = "2024-12"
open_weights = true

[limit]
context = 1_048_576
output = 131_072

[modalities]
input = ["text"]
output = ["text"]

[[weights]]
label = "Hugging Face"
url = "https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"

[[benchmarks]]
name = "SWE-Bench Verified"
score = 78.9
metric = "resolved"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro"

[[benchmarks]]
name = "SWE-Bench Pro"
score = 57.2
metric = "resolve rate"
source = "https://mimo.xiaomi.com/mimo-v2-5-pro/"
date = "2026-04-22"

[[benchmarks]]
name = "GPQA Diamond"
score = 86.6
metric = "accuracy"
source = "https://mimo.xiaomi.com/mimo-v2-5-pro/"
date = "2026-04-22"

[[benchmarks]]
name = "DeepSWE"
score = 19
metric = "score"
version = "1.1"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "ProgramBench"
score = 12.5
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "MiMo Code Bench"
score = 40.4
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "AutomationBench"
score = 16
metric = "score"
version = "1.0.6"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "Toolathlon-Verified"
score = 49.1
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "GDPval-AA"
score = 1107
metric = "Elo"
version = "2.1"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "Agents' Last Exam"
score = 13.2
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "Terminal-Bench"
score = 1.5
metric = "score"
version = "4.0"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "Terminal-Bench"
score = 65.2
metric = "score"
version = "2.1"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "JobBench"
score = 25
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "CyberGym"
score = 40
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "MiMo Cyber Bench"
score = 0
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "ExploitGym"
score = 0.2
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "ExploitBench"
score = 16.6
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"

[[benchmarks]]
name = "SEC-Bench Pro"
score = 17.7
metric = "score"
variant = "MiMo-V2.5 Pro comparison column"
source = "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL"