# Microsoft Phi 4

As of October 21, 2026, Microsoft Phi 4 ranks #201 of 213 models overall with 5.9 on the Artificial Analysis Intelligence Index. Its strongest category is knowledge, where it ranks #118 of 147. Microsoft Phi 4 costs $0.07 per million input tokens and $0.14 per million output tokens. It was released on January 10, 2025.

Page: https://llmbenchmark.ai/models/microsoft-phi-4
Data from October 21, 2026. Every figure below names the board it was read from and the day.

- Maker: Microsoft
- Released: January 10, 2025
- Price per million input tokens: $0.07
- Price per million output tokens: $0.14
- Context window: 16K tokens
- Overall rank: 201 of 213, by the Artificial Analysis Intelligence Index
- Artificial Analysis Intelligence Index: 5.9

Benchmark results (45 of 190):

- AA Intelligence Index: 5.9 (source: Artificial Analysis (via OpenRouter), read October 21, 2026)
- AA Math Index: 18.0 (source: Artificial Analysis, read October 21, 2026)
- GPQA Diamond (Artificial Analysis run): 57.5% (source: Artificial Analysis, read October 21, 2026)
- Humanity's Last Exam (Artificial Analysis run): 3.8% (source: Artificial Analysis, read October 21, 2026)
- MMLU Pro (Artificial Analysis run): 71.4% (source: Artificial Analysis, read October 21, 2026)
- LiveCodeBench (Artificial Analysis run): 23.1% (source: Artificial Analysis, read October 21, 2026)
- Terminal-Bench Hard (Artificial Analysis run): 3.8% (source: Artificial Analysis, read October 21, 2026)
- IFBench: 23.5% (source: Artificial Analysis, read October 21, 2026)
- Long context reasoning (AA-LCR): 0.0% (source: Artificial Analysis, read October 21, 2026)
- τ²-Bench Telecom: 0.0% (source: Artificial Analysis, read October 21, 2026)
- AIME 2025 (Artificial Analysis run): 18.0% (source: Artificial Analysis, read October 21, 2026)
- BALROG: 11.6% (source: Epoch AI benchmarking hub, read October 21, 2026)
- GPQA Diamond (Epoch AI run): 56.1% (source: Epoch AI benchmarking hub, read October 21, 2026)
- OTIS Mock AIME: 13.8% (source: Epoch AI benchmarking hub, read October 21, 2026)
- MATH level 5: 64.9% (source: Epoch AI benchmarking hub, read October 21, 2026)
- MMLU: 84.8% (source: Epoch AI benchmarking hub, read October 21, 2026)
- Creative writing (Lech Mazur): 62.6% (source: Epoch AI benchmarking hub, read October 21, 2026)
- Epoch Capabilities Index: 130.5 (source: Epoch AI benchmarking hub, read October 21, 2026)
- LMArena Text: 1217 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: coding: 1231 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: hard prompts: 1220 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: math: 1246 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: instruction following: 1201 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: creative writing: 1181 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: English: 1231 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: non-English: 1197 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: Chinese: 1212 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: German: 1222 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: Japanese: 1158 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: Korean: 1151 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: French: 1224 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: Spanish: 1234 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: Russian: 1209 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: multi-turn: 1206 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: longer queries: 1217 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: expert: 1203 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: hard prompts in English: 1230 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: software and IT: 1228 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: mathematical work: 1254 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: science: 1206 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: medicine and healthcare: 1202 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: legal and government: 1241 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: business and finance: 1199 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: writing and language: 1198 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)
- LMArena Text: entertainment and media: 1162 Elo (source: LMArena (Hugging Face dataset), read October 21, 2026)