ModelsCompareBest forBenchmarksStatusPricingAPI

Phi-4

Microsoft·Pending·Released 2024-12-12text
Context16.384k
Max output16.384k
Input / 1M-
Output / 1M-
Cached in-
Benchmarks11

Benchmarks

Cited public sources · rank is exact-benchmark, apples-to-apples

GPQA Diamond (Epoch AI run)#72/7756.1% ↗
OTIS Mock AIME 2024-2025 (Epoch AI run)#72/7613.8% ↗
LMArena Elo#60/621216.6 ↗
ArenaHard75.4% ↗
GPQA Diamond#63/7056.1% ↗
HumanEval#2/282.6% ↗
HumanEval+#4/482.8% ↗
IFEval#8/863% ↗
MATH#1/380.4% ↗
MMLU#10/1484.8% ↗
MMLU-Pro#24/2970.4% ↗

Compare

vs GPT-6 Astravs GPT-5.6 Solvs GPT-5.6 Terravs GPT-5.6 Lunavs Claude Fable 5.1vs Claude Opus 5vs Claude Sonnet 5vs Claude Haiku 4.5

More from Microsoft

MAI-Thinking-1Phi-4 ReasoningPhi-4 Mini ReasoningPhi-4 Multimodal Instruct

Access Phi-4 and every other model through one endpoint with automatic failover: Respan gateway.