DeepInfra models
Low-cost serverless inference for open models with transparent per-token rates.
Llama 3.3 70B Turbo (DeepInfra)
$0.1 / $0.32 per 1M tokens
Llama 3.1 8B Turbo (DeepInfra)
$0.02 / $0.04 per 1M tokens
DeepSeek V4 Flash (DeepInfra)
$0.09 / $0.18 per 1M tokens
DeepSeek V4 Pro (DeepInfra)
$1.30 / $2.60 per 1M tokens
Llama 4 Scout (DeepInfra)
$0.1 / $0.3 per 1M tokens
Llama 4 Maverick (DeepInfra)
$0.2 / $0.8 per 1M tokens
Qwen3 32B (DeepInfra)
$0.08 / $0.28 per 1M tokens
Mistral Nemo (DeepInfra)
$0.019 / $0.03 per 1M tokens
Gemma 4 31B Turbo (DeepInfra)
$0.09 / $0.34 per 1M tokens
DeepInfra
Low-cost serverless inference for open models with transparent per-token rates.
There are 9 published models here. On a 70/30 mix, combined list rates run from $0.0223 to $1.69 per 1 million tokens, with a median of $0.16. The largest listed context window in this group is 1,024,000 tokens. The most recent price check in this group is 2026-08-14.
Official DeepInfra pricing page
DeepInfra list rates
USD per million tokens. The 70/30 mix is a planning default — change it on the model page if your traffic is output-heavy.
| Model | Input / 1M | Output / 1M | Mixed 70/30 | Context | Verified |
|---|---|---|---|---|---|
| Mistral Nemo (DeepInfra) | $0.019 | $0.03 | $0.0223 | 128,000 | |
| Llama 3.1 8B Turbo (DeepInfra) | $0.02 | $0.04 | $0.026 | 128,000 | |
| DeepSeek V4 Flash (DeepInfra) | $0.09 | $0.18 | $0.117 | 1,024,000 | |
| Qwen3 32B (DeepInfra) | $0.08 | $0.28 | $0.14 | 40,000 | |
| Llama 4 Scout (DeepInfra) | $0.1 | $0.3 | $0.16 | 320,000 | |
| Gemma 4 31B Turbo (DeepInfra) | $0.09 | $0.34 | $0.165 | 256,000 | |
| Llama 3.3 70B Turbo (DeepInfra) | $0.1 | $0.32 | $0.166 | 128,000 | |
| Llama 4 Maverick (DeepInfra) | $0.2 | $0.8 | $0.38 | 1,024,000 | |
| DeepSeek V4 Pro (DeepInfra) | $1.30 | $2.60 | $1.69 | 1,024,000 |
Questions about DeepInfra rates
Do these DeepInfra figures include discounts and tax?
No. They are public list rates. Commits, credits, regions, tax, and commercial discounts are not in the number — check the source linked on each model.
When is the more expensive DeepInfra tier worth the extra?
DeepSeek V4 Pro (DeepInfra) sits near $1.69 per 1M tokens on a 70/30 mix, versus $0.0223 on Mistral Nemo (DeepInfra). Use the high tier when retries or long reasoning actually fail on the cheap one.