Gemma 2 27B vs Phi-3.5-MoE-instruct Comparison

Comparing Gemma 2 27B and Phi-3.5-MoE-instruct across benchmarks, pricing, and capabilities.

Performance Benchmarks

Comparative analysis across standard metrics

11 benchmarks

Gemma 2 27B outperforms in 3 benchmarks (BoolQ, HellaSwag, Winogrande), while Phi-3.5-MoE-instruct is better at 8 benchmarks (ARC-C, GSM8k, HumanEval, MATH, MBPP, MMLU, PIQA, Social IQa).

Phi-3.5-MoE-instruct shows notably better performance in the majority of benchmarks.

Thu Mar 19 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

Cost data unavailable.

Lowest available price from all providers
Thu Mar 19 2026 • llm-stats.com
Google
Gemma 2 27B
Input tokens$0.00
Output tokens$0.00
Best providerUnknown Organization
Microsoft
Phi-3.5-MoE-instruct
Input tokens$0.00
Output tokens$0.00
Best providerUnknown Organization
Notice missing or incorrect data?Start an Issue

Model Size

Parameter count comparison

32.8B diff

Phi-3.5-MoE-instruct has 32.8B more parameters than Gemma 2 27B, making it 120.6% larger.

Google
Gemma 2 27B
27.2Bparameters
Microsoft
Phi-3.5-MoE-instruct
60.0Bparameters
27.2B
Gemma 2 27B
60.0B
Phi-3.5-MoE-instruct

License

Usage and distribution terms

Gemma 2 27B is licensed under Gemma, while Phi-3.5-MoE-instruct uses MIT.

License differences may affect how you can use these models in commercial or open-source projects.

Gemma 2 27B

Gemma

Open weights

Phi-3.5-MoE-instruct

MIT

Open weights

Release Timeline

When each model was launched

Gemma 2 27B was released on 2024-06-27, while Phi-3.5-MoE-instruct was released on 2024-08-23.

Phi-3.5-MoE-instruct is 2 months newer than Gemma 2 27B.

Gemma 2 27B

Jun 27, 2024

1.7 years ago

Phi-3.5-MoE-instruct

Aug 23, 2024

1.6 years ago

1mo newer

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Key Takeaways

Higher BoolQ score (84.8% vs 84.6%)
Higher HellaSwag score (86.4% vs 83.8%)
Higher Winogrande score (83.7% vs 81.3%)
Higher ARC-C score (91.0% vs 71.4%)
Higher GSM8k score (88.7% vs 74.0%)
Higher HumanEval score (70.7% vs 51.8%)
Higher MATH score (59.5% vs 42.3%)
Higher MBPP score (80.8% vs 62.6%)
Higher MMLU score (78.9% vs 75.2%)
Higher PIQA score (88.6% vs 83.2%)
Higher Social IQa score (78.0% vs 53.7%)

Detailed Comparison

AI Model Comparison Table
Feature
Google
Gemma 2 27B
Microsoft
Phi-3.5-MoE-instruct