This model is deprecated. We only continue performance benchmarking for the default 10k input token workload. Results for other workloads are historical and no longer updated.
Google has launched a newer model, Gemma 4 31B. We suggest considering it instead.
For more information, see comparison of Gemma 4 31B to other models and API provider benchmarks for Gemma 4 31B.
Gemma 3 27B Instruct API Provider Benchmarking & Analysis
This analysis is intended to support you in choosing the best API provider of Gemma 3 27B Instruct for your use-case.
Fastest
Amazon
Nebius (FP8)
Parasail
Novita
DeepInfraOutput speed
Total 6 providers
Lowest Latency
AmazonNovitaDeepInfraParasailNebius (FP8)Time to first token
Total 6 providers
Lowest Price
Google (AI Studio)DeepInfraParasailNebius (FP8)NovitaBlended price (per 1M tokens)
Total 6 providers
Gemma 3 27B is available through 6 API providers, each offering different performance characteristics and pricing. Below is a comparison of the key metrics across providers.
- For output speed, the top providers are Amazon (68.5 t/s), Nebius (FP8) (64.5 t/s), and Parasail (45.4 t/s). Speed varies significantly across providers, with a 314% difference between the fastest and slowest.
- For latency, Amazon (1.45s), Novita (1.58s), and DeepInfra (1.63s) offer the lowest time to first token.
- For pricing, Google (AI Studio) (0.00), DeepInfra (0.09), and Parasail (0.09) offer the lowest blended prices per 1M tokens.
- Amazon offers the best performance with both the highest speed and lowest latency. For cost optimization, Google (AI Studio) provides the most competitive pricing.
Highlights
Pricing
Pricing: Cache Hit, Input, and Output
Pricing: Blended Price
Pricing: Cache Discount
Output Speed vs. Price
Speed
Measured by Output Speed (tokens per second)
Output Speed: Gemma 3 27B Instruct
Latency vs. Output Speed
Latency
Measured by Time (seconds) to First Token
Time to First Token
End-to-End Response Time
Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed
End-to-End Response Time
Key Comparison Metrics & API Features
Frequently Asked Questions
Common questions about Gemma 3 27B Instruct providers
Gemma 3 27B Instruct is available through 6 API providers: Amazon, DeepInfra, Nebius (FP8), Google (AI Studio), Parasail, and Novita. Each provider offers different performance characteristics and pricing.
Gemma 3 27B Instruct is currently available through 6 API providers that we benchmark and track.
The fastest providers for Gemma 3 27B Instruct by output speed are Amazon (68.5 t/s), Nebius (FP8) (64.5 t/s), and Parasail (45.4 t/s). Output speed measures how quickly tokens are generated after the model starts responding.
The most affordable providers for Gemma 3 27B Instruct by blended price are Google (AI Studio) ($0.00 per 1M tokens), DeepInfra ($0.09 per 1M tokens), and Parasail ($0.09 per 1M tokens). Blended price uses a 7:2:1 cache hit/input/output token ratio.
The providers with the lowest input token pricing for Gemma 3 27B Instruct are Google (AI Studio) ($0.00 per 1M input tokens), DeepInfra ($0.08 per 1M input tokens), and Parasail ($0.08 per 1M input tokens).
The providers with the lowest output token pricing for Gemma 3 27B Instruct are Google (AI Studio) ($0.00 per 1M output tokens), DeepInfra ($0.16 per 1M output tokens), and Novita ($0.20 per 1M output tokens).
4 of 6 providers support JSON mode for Gemma 3 27B Instruct: DeepInfra, Nebius (FP8), Parasail, and Novita.
4 of 6 providers support function calling for Gemma 3 27B Instruct: Amazon, DeepInfra, Nebius (FP8), and Parasail.
For Gemma 3 27B Instruct, Amazon offers the best performance with highest speed and lowest latency. For cost optimization, Google (AI Studio) provides the most competitive pricing.
When choosing a provider for Gemma 3 27B Instruct, consider: output speed (for throughput-intensive tasks), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and API features like JSON mode or function calling.
Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.
For information about Gemma 3 27B Instruct's intelligence, capabilities, modalities, and how it compares to other models, see the model overview page. View model overview