Compare · ModelsLive · 2 picked · head to head
Grok 4.5 vs Qwen2.5 72B Instruct
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Grok 4.5 wins on 6/6 benchmarks
Grok 4.5 wins 6 of 6 shared benchmarks. Leads in arena · general · knowledge.
Category leads
arena·Grok 4.5general·Grok 4.5knowledge·Grok 4.5math·Grok 4.5coding·Grok 4.5
Hype vs Reality
Attention vs performance
Grok 4.5
#104 by perf·#14 by attention
Qwen2.5 72B Instruct
#141 by perf·#1 by attention
Best value
Qwen2.5 72B Instruct
9.7x better value than Grok 4.5
Grok 4.5
13.3 pts/$
$4.00/M
Qwen2.5 72B Instruct
129.7 pts/$
$0.38/M
Vendor risk
Who is behind the model
xAI
$250.0B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
6 benchmarks · 2 models
Grok 4.5Qwen2.5 72B Instruct
Chatbot Arena Elo · Overall
Grok 4.5 leads by +162.9
Grok 4.5
1465.8
Qwen2.5 72B Instruct
1302.8
Dtbench
Grok 4.5 leads by +56.0
Grok 4.5
94.2
Qwen2.5 72B Instruct
38.2
GPQA diamond
Grok 4.5 leads by +61.5
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Grok 4.5
95.0
Qwen2.5 72B Instruct
33.5
Lmca
Grok 4.5 leads by +37.4
Grok 4.5
53.2
Qwen2.5 72B Instruct
15.8
OTIS Mock AIME 2024-2025
Grok 4.5 leads by +89.8
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Grok 4.5
97.8
Qwen2.5 72B Instruct
8.0
WeirdML
Grok 4.5 leads by +30.5
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Grok 4.5
46.4
Qwen2.5 72B Instruct
16.0
Full benchmark table
| Benchmark | Grok 4.5 | Qwen2.5 72B Instruct |
|---|---|---|
Chatbot Arena Elo · Overall | 1465.8 | 1302.8 |
Dtbench | 94.2 | 38.2 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 95.0 | 33.5 |
Lmca | 53.2 | 15.8 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 97.8 | 8.0 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 46.4 | 16.0 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $2.00 | $6.00 | 500K tokens (~250 books) | $30.00 | |
| $0.36 | $0.40 | 33K tokens (~16 books) | $3.70 |