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

Microsoft: phi-4

6 benchmarks
CompanyMicrosoft
Country of originUSA

Leaderboard position

#875
Overall rank (of 1121 models)
283
Elo rating
6
Benchmarks
Top 78.1%
Percentile
#12
In Microsoft (of 19 models)
Above: Kling AI: kling-2.6-pro #874 Below: Qwen: qwen2-72b-instruct #876

Benchmark highlights

Best benchmark
LEXam Mcq 4 Choices 40.66
Worst benchmark
CowQA-Rapid 0

Nearest competitor from another company

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

appeared as nearest in 1 of 6 benchmarks

Anthropic: claude-3-sonnet
1/6 16.7%
Ilya Gusev: saiga_llama3_8b_v7
1/6 16.7%
OpenAI: gpt-oss-20b
1/6 16.7%
Qwen: qwen3-32b
1/6 16.7%
Meta Llama: Llama 3.3 70B Instruct
1/6 16.7%
Qwen: Qwen3.5 0.8B
1/6 16.7%
Model Votes % of benchmarks
Anthropic: claude-3-sonnet 1 16.7%
Ilya Gusev: saiga_llama3_8b_v7 1 16.7%
OpenAI: gpt-oss-20b 1 16.7%
Qwen: qwen3-32b 1 16.7%
Meta Llama: Llama 3.3 70B Instruct 1 16.7%
Qwen: Qwen3.5 0.8B 1 16.7%

Benchmarks

LEXam Mcq 4 Choices 40.66 LEXam Open Question 38.54 SuperGPQA 37.65 Artificial Analysis Intelligence Index v4.0 (Maximum Reasoning) 10.4 Artificial Analysis Intelligence Index v4.1 (Maximum Reasoning) 5 CowQA-Rapid 0

Predicted scores

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

Benchmark Predicted Confidence Based on
GPQA Diamond ~45.3 61% Qwen: Qwen3 8B, Google: Gemma 3 12B Instruct, Meta Llama: Llama 3.3 70B Instruct
MATH-500 ~76.47 60% Qwen: Qwen3 8B, OpenBMB: minicpm5-1b
AIME 2026 ~62.44 60% OpenBMB: minicpm5-1b, Qwen: Qwen3.5 9B, DeepSeek: deepseek-v3.2
HMMT Feb 2026 ~53.91 60% OpenBMB: minicpm5-1b, Qwen: Qwen3.5 9B, DeepSeek: deepseek-v3.2
Humanity's Last Exam (no tools) ~4 60% Qwen: Qwen3 8B, Meta Llama: Llama 3.3 70B Instruct, OpenAI: gpt-oss-20b
IFEval ~74.92 60% Qwen: Qwen3 8B, Google: Gemma 3 12B Instruct, Qwen: Qwen3.5 0.8B
Terminal-Bench v2.1 ~8.6 60% DeepSeek: DeepSeek V4 Flash, SpaceXAI: Grok 4.3
MMMU-Pro ~42.22 60% Qwen: Qwen3.5 27B, Google: Gemini 3.5 Flash, Qwen: qwen3.6-27b
MATH ~36.78 60% OpenAI: o3-mini
BBH ~52.65 60% OpenBMB: minicpm5-1b
ARC-Easy ~67.15 60% Qwen: Qwen3.5 0.8B, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 9B
HellaSwag ~64.2 60% Qwen: Qwen3.5 0.8B, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 9B
PIQA ~76.62 60% Qwen: Qwen3.5 0.8B
Artificial Analysis Intelligence Index v4.3.2 ~5.08 60% Qwen: Qwen3.5 0.8B, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 9B
CowQA ~50.31 60% OpenAI: GPT-5 Nano, Qwen: Qwen3.5 9B, Xiaomi: mimo-v2-flash
BrowseComp ~24.69 60% Qwen: Qwen3.5 27B, Z.ai: GLM 5.1, Anthropic: Claude Opus 4.7
GPQA ~32.52 60% Qwen: Qwen3 8B, Qwen: Qwen3.5 0.8B, DeepSeek: deepseek-v3.2
ParseBench ~33.48 60% Qwen: Qwen3.5 0.8B, Qwen: Qwen3.5 9B
Terminal-Bench v2.0 ~1.39 60% DeepSeek: deepseek-v3.2, Qwen: qwen3.5-122b-a10b, Qwen: Qwen3.5 27B
ScreenSpot-Pro ~39.34 60% Qwen: Qwen3.5 0.8B, Qwen: Qwen3.5 9B, Qwen: qwen3.5-122b-a10b
Text Arena (Snapshot from 20 june 2026) ~1262.44 60% Meta Llama: Llama 3.1 405B Instruct, Google: Gemma 3 12B Instruct, OpenAI: gpt-oss-20b
WinoGrande ~65.35 60% Qwen: Qwen3.5 0.8B, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 9B
TruthfulQA ~37.25 60% Qwen: Qwen3.5 0.8B, OpenBMB: minicpm5-1b, Qwen: Qwen3.5 9B
MMLU-Pro ~49.52 59% Qwen: Qwen3.5 0.8B, DeepSeek: deepseek-v3, OpenBMB: minicpm5-1b
SWE-bench Pro ~5.24 59% Meta Llama: Llama 3.1 405B Instruct, Qwen: qwen3-235b-a22b, Google: Gemma 3 27B Instruct