Gemma 4 31B (Non-reasoning) API Provider Benchmarking & Analysis
This analysis is intended to support you in choosing the best API provider of Gemma 4 31B (Non-reasoning) for your use-case.
Fastest
Modular (NVFP4)
SambaNova
FriendliAI
SiliconFlow (FP8)
CoreWeaveOutput speed
Total 11 providers
Lowest Latency
Modular (NVFP4)
Crusoe
DeepInfra (FP8)
Parasail
Together AI (FP8)Time to first token
Total 11 providers
Lowest Price
ParasailCoreWeaveDeepInfra (FP8)SiliconFlow (FP8)FriendliAIBlended price (per 1M tokens)
Total 11 providers
Gemma 4 31B (Non-reasoning) is available through 11 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 Modular (NVFP4) (208.1 t/s), SambaNova (203.1 t/s), and FriendliAI (107.7 t/s). Speed varies significantly across providers, with a 1348% difference between the fastest and slowest.
- For latency, Modular (NVFP4) (0.43s), Crusoe (1.18s), and DeepInfra (FP8) (1.46s) offer the lowest time to first token.
- For pricing, Parasail (0.11), CoreWeave (0.12), and DeepInfra (FP8) (0.15) offer the lowest blended prices per 1M tokens. Prices vary up to 9.3x across providers.
- Modular (NVFP4) offers the best performance with both the highest speed and lowest latency. For cost optimization, Parasail provides the most competitive pricing.
Highlights
Update: Default performance benchmarking workload has updated to 10k input tokens to better reflect production use cases. You can still select different workloads above.
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 4 31B (Non-reasoning)
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
| Modular | 262k | Open | -- | 208 | 0.43 | 2.83 | -- | ||
| Parasail | 262k | Open | -- | 36 | 1.50 | 15.43 | -- | ||
 Cerebras | 131k | Open | -- | -- | -- | -- | -- | ||
| FriendliAI | 262k | Open | -- | 108 | 2.17 | 6.81 | -- | ||
| Together AI | 262k | Open | -- | 14 | 1.53 | 36.31 | -- | ||
| SiliconFlow | 262k | Open | -- | 57 | 3.45 | 12.25 | -- | ||
| DeepInfra | 262k | Open | -- | 15 | 1.46 | 35.70 | -- | ||
| CoreWeave | 262k | Open | -- | 56 | 2.10 | 11.10 | -- | ||
| SambaNova | 131k | Open | -- | 203 | 2.67 | 5.14 | -- | ||
| Crusoe | 262k | Open | -- | 18 | 1.18 | 29.41 | -- | ||
 Novita | 262k | Open | -- | 15 | 2.01 | 35.93 | -- | ||
Frequently Asked Questions
Common questions about Gemma 4 31B (Non-reasoning) providers
Gemma 4 31B (Non-reasoning) is available through 11 API providers: Modular (NVFP4), Parasail, Cerebras, FriendliAI, Together AI (FP8), SiliconFlow (FP8), DeepInfra (FP8), CoreWeave, SambaNova, Crusoe, and Novita. Each provider offers different performance characteristics and pricing.
Gemma 4 31B (Non-reasoning) is currently available through 11 API providers that we benchmark and track.
The fastest providers for Gemma 4 31B (Non-reasoning) by output speed are Modular (NVFP4) (208.1 t/s), SambaNova (203.1 t/s), and FriendliAI (107.7 t/s). Output speed measures how quickly tokens are generated after the model starts responding.
The providers with the lowest time to first token for Gemma 4 31B (Non-reasoning) are Modular (NVFP4) (0.43s), Crusoe (1.18s), and DeepInfra (FP8) (1.46s). Lower latency means faster initial response time.
The most affordable providers for Gemma 4 31B (Non-reasoning) by blended price are Parasail ($0.11 per 1M tokens), CoreWeave ($0.12 per 1M tokens), and DeepInfra (FP8) ($0.15 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 4 31B (Non-reasoning) are CoreWeave ($0.10 per 1M input tokens), SiliconFlow (FP8) ($0.13 per 1M input tokens), and DeepInfra (FP8) ($0.13 per 1M input tokens).
The providers with the lowest output token pricing for Gemma 4 31B (Non-reasoning) are CoreWeave ($0.34 per 1M output tokens), DeepInfra (FP8) ($0.38 per 1M output tokens), and Parasail ($0.40 per 1M output tokens).
Output speed for Gemma 4 31B (Non-reasoning) varies significantly across providers. Modular (NVFP4) is the fastest at 208.1 t/s, which is 14.5x faster than Together AI (FP8) at 14.4 t/s.
All 11 providers of Gemma 4 31B (Non-reasoning) support JSON mode for structured output.
All 11 providers of Gemma 4 31B (Non-reasoning) support function calling (tool use).
For Gemma 4 31B (Non-reasoning), Modular (NVFP4) offers the best performance with highest speed and lowest latency. For cost optimization, Parasail provides the most competitive pricing.
When choosing a provider for Gemma 4 31B (Non-reasoning), 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 4 31B (Non-reasoning)'s intelligence, capabilities, modalities, and how it compares to other models, see the model overview page. View model overview