Skip to content

Mistral Large 2 vs Phi-4

Mistral
Mistral Large 2

Mistral AI · 123B params · Quality: 82

Microsoft
Phi-4

Microsoft · 14.7B params · Quality: 83

Architecture Comparison

SpecMistral Large 2Phi-4
TypeDENSEDENSE
Total Parameters123B14.7B
Active Parameters123B14.7B
Layers8840
Hidden Dimension12,2885,120
Attention Heads9640
KV Heads810
Context Length131,07216,384
Precision (default)BF16BF16

Memory Requirements

PrecisionMistral Large 2Phi-4
BF16 Weights246.0 GB29.4 GB
FP8 Weights123.0 GB14.7 GB
INT4 Weights61.5 GB7.3 GB
KV-Cache / Token360448 B204800 B
Activation Estimate3.50 GB1.50 GB

Minimum GPUs Needed (BF16)

H100 SXM4 GPUs1 GPU
L40S7 GPUs1 GPU

Quality Benchmarks

BenchmarkMistral Large 2Phi-4
Overall8283
MMLU84.084.8
HumanEval53.067.0
GSM8K91.293.0
MT-Bench84.085.0

Mistral Large 2

MMLU
84.0
HumanEval
53.0
GSM8K
91.2
MT-Bench
84.0

Phi-4

MMLU
84.8
HumanEval
67.0
GSM8K
93.0
MT-Bench
85.0

Capabilities

FeatureMistral Large 2Phi-4
Tool Use✓ Yes✓ Yes
Vision✗ No✗ No
Code✓ Yes✓ Yes
Math✓ Yes✓ Yes
Reasoning✗ No✓ Yes
Multilingual✓ Yes✓ Yes
Structured Output✓ Yes✓ Yes

API Pricing Comparison

Cheapest Output (Mistral Large 2)

$2.50/M

Input: $2.50/M

Cheapest Output (Phi-4)

$0.14/M

Input: $0.07/M

ProviderMistral Large 2 In $/MOut $/MPhi-4 In $/MOut $/M
azure$0.07$0.14
together$2.50$2.50$0.20$0.20
mistral$2.00$6.00

Recommendation Summary

  • Phi-4 scores higher on overall quality (83 vs 82).
  • Phi-4 is cheaper per output token ($0.14/M vs $2.50/M).
  • Phi-4 has a smaller memory footprint (29.4 GB vs 246.0 GB BF16), making it easier to deploy on fewer GPUs.
  • Mistral Large 2 supports a longer context window (131,072 vs 16,384 tokens).
  • Phi-4 is stronger at code generation (HumanEval: 67.0 vs 53.0).
  • Phi-4 is better at math reasoning (GSM8K: 93.0 vs 91.2).

Compare Other Models