Same open weights, served by 2 providers. Weights: google/gemma-4-31B-it.
| Provider | Their model id | Serving | Input | Output | Cached in | Blended 3:1 | Source | |
|---|---|---|---|---|---|---|---|---|
DeepInfra | google/gemma-4-31B-it-turbo | flexFP40.8× standard | $0.072 | $0.272 | $0.04 | $0.122 | Verified 5h ago · Changed 7h ago | |
DeepInfra | google/gemma-4-31B-it-turbo | standardFP4 | $0.09 | $0.34 | $0.05 | $0.153 | Verified 5h ago · Changed 7h ago | |
DeepInfra | google/gemma-4-31B-it | flexFP80.8× standard | $0.16 | $0.32 | — | $0.20 | Verified 5h ago · Changed 7h ago | |
AWS Bedrock | google.gemma-4-31b | standardQuantization not disclosed | $0.14 | $0.40 | — | $0.205 | Verified 7h ago · Changed 7h ago | |
DeepInfra | google/gemma-4-31B-it-turbo | priorityFP41.5× standard | $0.135 | $0.51 | $0.075 | $0.229 | Verified 5h ago · Changed 7h ago | |
DeepInfra | google/gemma-4-31B-it | standardFP8 | $0.20 | $0.40 | — | $0.25 | Verified 5h ago · Changed 7h ago | |
DeepInfra | google/gemma-4-31B-it | priorityFP81.5× standard | $0.30 | $0.60 | — | $0.375 | Verified 5h ago · Changed 7h ago |
Same weights don't always mean identical outputs: quantization, context limits and serving stacks can change quality, especially for tool calling. Prices are USD per 1M tokens; hyperscaler prices are for the reference region shown.
| DeepInfra fp4 | $15.8/mo |
| AWS Bedrock | $22/mo |
| DeepInfra fp8 | $28/mo |
Cache hits use each provider's cached-input price where it has one. Cache write costs aren't included.