Gemma 3 27B vs Mistral Large 2
Architecture Comparison
SpecGemma 3 27BMistral Large 2
TypeDENSEDENSE
Total Parameters27B123B
Active Parameters27B123B
Layers6288
Hidden Dimension3,58412,288
Attention Heads3296
KV Heads168
Context Length131,072131,072
Precision (default)BF16BF16
Memory Requirements
PrecisionGemma 3 27BMistral Large 2
BF16 Weights54.0 GB246.0 GB
FP8 Weights27.0 GB123.0 GB
INT4 Weights13.5 GB61.5 GB
KV-Cache / Token507904 B360448 B
Activation Estimate1.50 GB3.50 GB
Minimum GPUs Needed (BF16)
H100 SXM1 GPU4 GPUs
L40S2 GPUs7 GPUs
Quality Benchmarks
BenchmarkGemma 3 27BMistral Large 2
Overall7682
MMLU78.084.0
HumanEval48.053.0
GSM8K85.091.2
MT-Bench82.084.0
Gemma 3 27B
MMLU
78.0
HumanEval
48.0
GSM8K
85.0
MT-Bench
82.0
Mistral Large 2
MMLU
84.0
HumanEval
53.0
GSM8K
91.2
MT-Bench
84.0
Capabilities
FeatureGemma 3 27BMistral Large 2
Tool Use✓ Yes✓ Yes
Vision✓ Yes✗ No
Code✓ Yes✓ Yes
Math✓ Yes✓ Yes
Reasoning✗ No✗ No
Multilingual✓ Yes✓ Yes
Structured Output✓ Yes✓ Yes
API Pricing Comparison
Cheapest Output (Gemma 3 27B)
$0.20/M
Input: $0.10/M
Cheapest Output (Mistral Large 2)
$2.50/M
Input: $2.50/M
| Provider | Gemma 3 27B In $/M | Out $/M | Mistral Large 2 In $/M | Out $/M |
|---|---|---|---|---|
| $0.10 | $0.20 | — | — | |
| together | $0.30 | $0.30 | $2.50 | $2.50 |
| mistral | — | — | $2.00 | $6.00 |
Recommendation Summary
- ‣Mistral Large 2 scores higher on overall quality (82 vs 76).
- ‣Gemma 3 27B is cheaper per output token ($0.20/M vs $2.50/M).
- ‣Gemma 3 27B has a smaller memory footprint (54.0 GB vs 246.0 GB BF16), making it easier to deploy on fewer GPUs.
- ‣Mistral Large 2 is stronger at code generation (HumanEval: 53.0 vs 48.0).
- ‣Mistral Large 2 is better at math reasoning (GSM8K: 91.2 vs 85.0).