Compare · ModelsLive · 2 picked · head to head
phi-3-medium 14B vs Qwen2.5 72B Instruct
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Qwen2.5 72B Instruct wins on 6/8 benchmarks
Qwen2.5 72B Instruct wins 6 of 8 shared benchmarks. Leads in knowledge · math.
Category leads
knowledge·Qwen2.5 72B Instructreasoning·phi-3-medium 14Bmath·Qwen2.5 72B Instruct
Hype vs Reality
Attention vs performance
phi-3-medium 14B
#48 by perf·no signal
Qwen2.5 72B Instruct
#82 by perf·no signal
Best value
Qwen2.5 72B Instruct
phi-3-medium 14B
—
no price
Qwen2.5 72B Instruct
140.0 pts/$
$0.38/M
Vendor risk
Who is behind the model
Microsoft
$3.00T·Big Tech
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
8 benchmarks · 2 models
phi-3-medium 14BQwen2.5 72B Instruct
ARC AI2
Qwen2.5 72B Instruct leads by +3.9
AI2 Reasoning Challenge · tests grade-school level science knowledge with multiple-choice questions requiring reasoning beyond simple retrieval.
phi-3-medium 14B
88.8
Qwen2.5 72B Instruct
92.7
BBH
phi-3-medium 14B leads by +2.1
BIG-Bench Hard · a curated subset of 23 challenging tasks from BIG-Bench where language models previously failed to outperform average humans.
phi-3-medium 14B
75.2
Qwen2.5 72B Instruct
73.1
GPQA diamond
Qwen2.5 72B Instruct leads by +28.8
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
phi-3-medium 14B
3.5
Qwen2.5 72B Instruct
32.2
HellaSwag
Qwen2.5 72B Instruct leads by +3.2
HellaSwag · tests commonsense reasoning by asking models to predict the most plausible continuation of everyday scenarios.
phi-3-medium 14B
76.5
Qwen2.5 72B Instruct
79.7
MATH level 5
Qwen2.5 72B Instruct leads by +45.6
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
phi-3-medium 14B
17.6
Qwen2.5 72B Instruct
63.2
MMLU
Qwen2.5 72B Instruct leads by +9.7
Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge.
phi-3-medium 14B
70.7
Qwen2.5 72B Instruct
80.4
TriviaQA
phi-3-medium 14B leads by +2.0
TriviaQA · reading comprehension benchmark with trivia questions, requiring models to find and reason over evidence from provided documents.
phi-3-medium 14B
73.9
Qwen2.5 72B Instruct
71.9
Winogrande
Qwen2.5 72B Instruct leads by +1.6
WinoGrande · large-scale commonsense reasoning benchmark where models must resolve ambiguous pronouns in carefully constructed sentence pairs.
phi-3-medium 14B
63.0
Qwen2.5 72B Instruct
64.6
Full benchmark table
| Benchmark | phi-3-medium 14B | Qwen2.5 72B Instruct |
|---|---|---|
ARC AI2 AI2 Reasoning Challenge · tests grade-school level science knowledge with multiple-choice questions requiring reasoning beyond simple retrieval. | 88.8 | 92.7 |
BBH BIG-Bench Hard · a curated subset of 23 challenging tasks from BIG-Bench where language models previously failed to outperform average humans. | 75.2 | 73.1 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 3.5 | 32.2 |
HellaSwag HellaSwag · tests commonsense reasoning by asking models to predict the most plausible continuation of everyday scenarios. | 76.5 | 79.7 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 17.6 | 63.2 |
MMLU Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge. | 70.7 | 80.4 |
TriviaQA TriviaQA · reading comprehension benchmark with trivia questions, requiring models to find and reason over evidence from provided documents. | 73.9 | 71.9 |
Winogrande WinoGrande · large-scale commonsense reasoning benchmark where models must resolve ambiguous pronouns in carefully constructed sentence pairs. | 63.0 | 64.6 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| — | — | — | — | |
| $0.36 | $0.40 | 33K tokens (~16 books) | $3.70 |