Start typing to search.
Ranks providers by estimated cost using a 1,000-input/500-output token ratio, shown per 1 million total tokens.
Weights & Biases currently ranks first at $0.196667 / 1M tokens via wandb/bf16.
| Rank | Provider / route | Estimated cost / 1M tokens | Speed measurements | Endpoint details |
|---|---|---|---|---|
| #1 | Provider / route Weights & Biases wandb/bf16 | Estimated cost / 1M tokens$0.196667 / 1M tokens | Speed measurements
| Endpoint details
|
| #2 | Provider / route Venice venice/bf16 | Estimated cost / 1M tokens$0.2 / 1M tokens | Speed measurements
| Endpoint details
|
| #3 | Provider / route Chutes chutes/fp4 | Estimated cost / 1M tokens$0.203333 / 1M tokens | Speed measurements
| Endpoint details
|
| #4 | Provider / route DeepInfra deepinfra/turbo | Estimated cost / 1M tokens$0.203333 / 1M tokens | Speed measurements
| Endpoint details
|
| #5 | Provider / route DeepInfra deepinfra/fp8 | Estimated cost / 1M tokens$0.213333 / 1M tokens | Speed measurements
| Endpoint details
|
| #6 | Provider / route SiliconFlow siliconflow/fp8 | Estimated cost / 1M tokens$0.22 / 1M tokens | Speed measurements
| Endpoint details
|
| #7 | Provider / route NovitaAI novita/bf16 | Estimated cost / 1M tokens$0.226667 / 1M tokens | Speed measurements
| Endpoint details
|
| #8 | Provider / route Parasail parasail/fp8 | Estimated cost / 1M tokens$0.233333 / 1M tokens | Speed measurements
| Endpoint details
|
| #9 | Provider / route Phala phala | Estimated cost / 1M tokens$0.253333 / 1M tokens | Speed measurements
| Endpoint details
|
| #10 | Provider / route ModelRun modelrun/fp4 | Estimated cost / 1M tokens$0.33 / 1M tokens | Speed measurements
| Endpoint details
|
| #11 | Provider / route Together together | Estimated cost / 1M tokens$0.473333 / 1M tokens | Speed measurements
| Endpoint details
|
| #12 | Provider / route Together together | Estimated cost / 1M tokens$0.583333 / 1M tokens | Speed measurements
| Endpoint details
|
| #13 | Provider / route SambaNova sambanova | Estimated cost / 1M tokens$0.636667 / 1M tokens | Speed measurements
| Endpoint details
|
| #14 | Provider / route Cerebras cerebras/fp16 | Estimated cost / 1M tokens$1.156667 / 1M tokens | Speed measurements
| Endpoint details
|
Endpoint data fetched .
Gemma 4 31B endpoint guide
14 of 14 Gemma 4 31B endpoints currently have the published data required for this ranking. Weights & Biases leads at $0.196667 / 1M tokens via wandb/bf16. The table keeps unranked routes visible so missing measurements do not look like missing provider availability.
The cheapest ranking uses a 1,000-input/500-output-token ratio and reports the blended result per 1 million total tokens. Because providers can price input and output differently, a workload with a different token ratio may produce a different order.
Venice currently ranks second at $0.2 / 1M tokens. Compare that gap with input and output price, response-start time, output speed, 30-minute uptime, context length, and the exact route before deciding whether first place is meaningful for your workload.
Gemma 4 31B has 14 published provider options, and performance data is matched to each exact OpenRouter routing tag. Different quantization, context, regional deployment, or provider configuration can change price and behavior even when the underlying model name is identical.
