Note: Overall leaderboard rankings may not reflect true model quality — individual benchmarks give a clearer picture. ARC-Challenge MMLU GPQA GSM8K Artificial Analysis Intelligence Index v4.1
← Back to leaderboard

Xiaomi: mimo-v2-flash

8 benchmarks
CompanyXiaomi 小米集团
Country of originChina
WeightsOpen weights
Parameter count310.0B

Leaderboard position

#410
Overall rank (of 1126 models)
1307
Elo rating
8
Benchmarks
Top 36.4%
Percentile
In Xiaomi (of 8 models)

Benchmark highlights

Nearest competitor from another company

The model from another company with the most similar benchmark scores.

appeared as nearest in 1 of 8 benchmarks

Benchmarks

Predicted scores

Estimated from 25 similar models and the global Elo rating. These are projections, not measured results.

Benchmark Predicted Confidence Based on
MMLU-Pro ~74.15 62% Qwen: Qwen3.5 35B A3B, MiniMax: MiniMax M2.1, Arcee AI: trinity-large-thinking
Humanity's Last Exam (no tools) ~19.95 61% OpenAI: gpt-oss-120b, Google: Gemma 4 31B Instruct, Qwen: Qwen3.5 27B
HMMT Feb 2026 ~73.8 61% Qwen: Qwen3.5 35B A3B, Qwen: Qwen3.5 9B, Qwen: Qwen3.5 27B
GPQA Diamond ~69.61 61% Arcee AI: trinity-large-thinking, OpenAI: gpt-oss-120b, Qwen: Qwen3.5 9B
IFBench ~61.56 61% Arcee AI: trinity-large-thinking, Inception: Mercury 2, Mistral AI: mistral-medium-3.5-128b
SWE-bench Verified ~54.37 61% Arcee AI: trinity-large-thinking, OpenAI: gpt-oss-120b, Mistral AI: mistral-medium-3.5-128b
SWE-bench Pro ~32.3 61% MiniMax: MiniMax M2.1, OpenAI: gpt-oss-120b, Qwen: Qwen3.5 27B
SuperGPQA ~52.03 61% OpenAI: GPT-5 Mini, OpenAI: gpt-oss-120b, SpaceXAI: Grok 4
Artificial Analysis Intelligence Index v4.1.1 (Maximum Reasoning) ~25.43 61% Anthropic: claude-4.5-haiku, Qwen: qwen3-max, OpenAI: gpt-oss-120b
Terminal-Bench v2.0 ~32.5 61% Qwen: Qwen3.5 35B A3B, MiniMax: MiniMax M2.1, Qwen: Qwen3.5 27B
ARC-Challenge ~82.54 60% Qwen: Qwen3.5 9B, Qwen: Qwen3.5 2B, OpenBMB: minicpm5-1b
HellaSwag ~93.77 60% Qwen: Qwen3.5 9B, Qwen: Qwen3.5 2B, OpenBMB: minicpm5-1b
PIQA ~84.39 60% Qwen: Qwen3.5 0.8B, Google: Gemma 3 270M Instruct
WinoGrande ~81.73 60% Qwen: Qwen3.5 9B, Qwen: Qwen3.5 2B, OpenBMB: minicpm5-1b
AIME 2025 ~81.3 60% Arcee AI: trinity-large-thinking, Inception: Mercury 2, Mistral AI: mistral-medium-3.5-128b
IFEval ~89.03 60% Qwen: Qwen3.5 9B, Qwen: Qwen3.5 27B, OpenBMB: minicpm5-1b
HumanEval ~95.9 60% Qwen: Qwen3 4B, OpenAI: gpt-4o, OpenAI: GPT-3.5 Turbo
MBPP ~93.7 60% Meta Llama: Llama 4 Scout Instruct, Qwen: Qwen3 4B, OpenAI: gpt-4o
MMLU ~79.17 60% Qwen: Qwen3.5 9B, DeepSeek: DeepSeek R1, Meta Llama: Llama 4 Scout Instruct
SciCode ~34.35 60% Inception: Mercury 2, DeepSeek: DeepSeek V4 Flash, Xiaomi: MiMo-V2.5
AA-LCR ~52.22 60% DeepSeek: DeepSeek V4 Flash, Xiaomi: MiMo-V2.5, SpaceXAI: Grok 4.3
Humanity's Last Exam (with tools) ~27.03 60% OpenAI: gpt-oss-120b, Google: Gemma 4 31B Instruct, Z.ai: GLM 5.1
-(CoderPPL) ~-1.69 60% Qwen: Qwen3.5 9B, OpenAI: gpt-oss-20b, Google: Gemma 4 E4B Instruct
LiveCodeBench v6 ~66.3 60% Qwen: Qwen3.5 9B, Google: Gemma 4 31B Instruct, Qwen: Qwen3.5 27B
MMMU-Pro ~62.51 60% Google: Gemma 4 31B Instruct, Qwen: Qwen3.5 27B, Qwen: qwen3.6-27b