Skip to content

Qwen 2.5 72B vs Gemma 3 27B

Alibaba
Qwen 2.5 72B

Alibaba · 72.7B params · Quality: 84

Google
Gemma 3 27B

Google · 27B params · Quality: 76

Architecture Comparison

SpecQwen 2.5 72BGemma 3 27B
TypeDENSEDENSE
Total Parameters72.7B27B
Active Parameters72.7B27B
Layers8062
Hidden Dimension8,1923,584
Attention Heads6432
KV Heads816
Context Length131,072131,072
Precision (default)BF16BF16

Memory Requirements

PrecisionQwen 2.5 72BGemma 3 27B
BF16 Weights145.4 GB54.0 GB
FP8 Weights72.7 GB27.0 GB
INT4 Weights36.4 GB13.5 GB
KV-Cache / Token327680 B507904 B
Activation Estimate2.50 GB1.50 GB

Minimum GPUs Needed (BF16)

H100 SXM3 GPUs1 GPU
L40S4 GPUs2 GPUs

Quality Benchmarks

BenchmarkQwen 2.5 72BGemma 3 27B
Overall8476
MMLU85.378.0
HumanEval56.048.0
GSM8K91.685.0
MT-Bench86.082.0

Qwen 2.5 72B

MMLU
85.3
HumanEval
56.0
GSM8K
91.6
MT-Bench
86.0

Gemma 3 27B

MMLU
78.0
HumanEval
48.0
GSM8K
85.0
MT-Bench
82.0

Capabilities

FeatureQwen 2.5 72BGemma 3 27B
Tool Use✓ Yes✓ Yes
Vision✗ No✓ Yes
Code✓ Yes✓ Yes
Math✓ Yes✓ Yes
Reasoning✗ No✗ No
Multilingual✓ Yes✓ Yes
Structured Output✓ Yes✓ Yes

API Pricing Comparison

Cheapest Output (Qwen 2.5 72B)

$0.90/M

Input: $0.90/M

Cheapest Output (Gemma 3 27B)

$0.20/M

Input: $0.10/M

ProviderQwen 2.5 72B In $/MOut $/MGemma 3 27B In $/MOut $/M
google$0.10$0.20
together$0.90$0.90$0.30$0.30
fireworks$0.90$0.90

Recommendation Summary

  • Qwen 2.5 72B scores higher on overall quality (84 vs 76).
  • Gemma 3 27B is cheaper per output token ($0.20/M vs $0.90/M).
  • Gemma 3 27B has a smaller memory footprint (54.0 GB vs 145.4 GB BF16), making it easier to deploy on fewer GPUs.
  • Qwen 2.5 72B is stronger at code generation (HumanEval: 56.0 vs 48.0).
  • Qwen 2.5 72B is better at math reasoning (GSM8K: 91.6 vs 85.0).

Compare Other Models