Compare · ModelsLive · 2 picked · head to head
GPT-4o (2024-11-20) vs Qwen2.5 72B Instruct
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-4o (2024-11-20) wins on 9/13 benchmarks
GPT-4o (2024-11-20) wins 9 of 13 shared benchmarks. Leads in coding · knowledge · general.
Category leads
coding·GPT-4o (2024-11-20)knowledge·GPT-4o (2024-11-20)general·GPT-4o (2024-11-20)math·Qwen2.5 72B Instructagentic·GPT-4o (2024-11-20)multimodal·Qwen2.5 72B Instruct
Hype vs Reality
Attention vs performance
GPT-4o (2024-11-20)
#255 by perf·no signal
Qwen2.5 72B Instruct
#142 by perf·#1 by attention
Best value
Qwen2.5 72B Instruct
25.0x better value than GPT-4o (2024-11-20)
GPT-4o (2024-11-20)
5.2 pts/$
$6.25/M
Qwen2.5 72B Instruct
129.7 pts/$
$0.38/M
Vendor risk
Who is behind the model
OpenAI
$840.0B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
13 benchmarks · 2 models
GPT-4o (2024-11-20)Qwen2.5 72B Instruct
Aider · Code Editing
GPT-4o (2024-11-20) leads by +6.0
GPT-4o (2024-11-20)
71.4
Qwen2.5 72B Instruct
65.4
Balrog
GPT-4o (2024-11-20) leads by +16.1
Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning.
GPT-4o (2024-11-20)
32.3
Qwen2.5 72B Instruct
16.2
Dtbench
GPT-4o (2024-11-20) leads by +2.7
GPT-4o (2024-11-20)
40.9
Qwen2.5 72B Instruct
38.2
GeoBench
GPT-4o (2024-11-20) leads by +9.0
GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding.
GPT-4o (2024-11-20)
71.0
Qwen2.5 72B Instruct
62.0
GPQA diamond
Qwen2.5 72B Instruct leads by +3.0
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
GPT-4o (2024-11-20)
30.5
Qwen2.5 72B Instruct
33.5
Lmca
GPT-4o (2024-11-20) leads by +3.7
GPT-4o (2024-11-20)
19.5
Qwen2.5 72B Instruct
15.8
MATH level 5
Qwen2.5 72B Instruct leads by +13.4
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
GPT-4o (2024-11-20)
49.8
Qwen2.5 72B Instruct
63.2
Metr Time Horizons
GPT-4o (2024-11-20) leads by +5.0
GPT-4o (2024-11-20)
40.8
Qwen2.5 72B Instruct
35.8
MMLU
GPT-4o (2024-11-20) leads by +3.7
Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge.
GPT-4o (2024-11-20)
84.1
Qwen2.5 72B Instruct
80.4
OTIS Mock AIME 2024-2025
Qwen2.5 72B Instruct leads by +1.8
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GPT-4o (2024-11-20)
6.2
Qwen2.5 72B Instruct
8.0
The Agent Company
GPT-4o (2024-11-20) leads by +2.9
The Agent Company · tests AI agents on realistic corporate tasks like email management, code review, data analysis, and cross-tool workflows.
GPT-4o (2024-11-20)
8.6
Qwen2.5 72B Instruct
5.7
VideoMME
Qwen2.5 72B Instruct leads by +2.1
VideoMME · multimodal benchmark testing video understanding across diverse domains, requiring temporal reasoning and cross-frame comprehension.
GPT-4o (2024-11-20)
62.5
Qwen2.5 72B Instruct
64.7
WeirdML
GPT-4o (2024-11-20) leads by +9.2
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
GPT-4o (2024-11-20)
25.1
Qwen2.5 72B Instruct
16.0
Full benchmark table
| Benchmark | GPT-4o (2024-11-20) | Qwen2.5 72B Instruct |
|---|---|---|
Aider · Code Editing | 71.4 | 65.4 |
Balrog Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning. | 32.3 | 16.2 |
Dtbench | 40.9 | 38.2 |
GeoBench GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding. | 71.0 | 62.0 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 30.5 | 33.5 |
Lmca | 19.5 | 15.8 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 49.8 | 63.2 |
Metr Time Horizons | 40.8 | 35.8 |
MMLU Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge. | 84.1 | 80.4 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 6.2 | 8.0 |
The Agent Company The Agent Company · tests AI agents on realistic corporate tasks like email management, code review, data analysis, and cross-tool workflows. | 8.6 | 5.7 |
VideoMME VideoMME · multimodal benchmark testing video understanding across diverse domains, requiring temporal reasoning and cross-frame comprehension. | 62.5 | 64.7 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 25.1 | 16.0 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $2.50 | $10.00 | 128K tokens (~64 books) | $43.75 | |
| $0.36 | $0.40 | 33K tokens (~16 books) | $3.70 |