Only published terms. A tier or mode the provider does not publish is written out, never interpolated from another one.
Standard · Any prompt size
| Component | PriceUSD / 1M tokens |
|---|---|
| Input (uncached) | 0.30 |
| Cached input (read) | 0.006 |
| Cache write | Free |
| Output | 1.20 |
Each route is a separate deployment of this offer with its own published prices. Only published prices are shown.
| Route | Input | Cached input | Output | Last checked |
|---|---|---|---|---|
| DeepSeek V4 Flash | 0.30 | 0.006 | 1.20 |
Every published revision of this offer, newest last in the chart. Earlier terms are kept; a correction adds a revision; a change of model version is marked as a break.
Only one set of terms is recorded, so there is no chart yet. The revision is listed below.
Input 0.30 USD / 1M tokensCached input (read) 0.006 USD / 1M tokensOutput 1.20 USD / 1M tokens
Other providers, tiers and subscriptions for the same model version. Prices differ by provider, region and conditions.
| Provider | Input | Output | Cached input |
|---|---|---|---|
| Vercel AI Gateway7 routes · from 0.06 | |||
| DeepSeek-V4-Flash, Vercel AI Gateway, Upstream provider: DeepInfra, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.06 | 0.18 | 0.015 |
| DeepSeek-V4-Flash, Vercel AI Gateway, Upstream provider: DigitalOceanUpstream provider — the company actually serving this route | 0.112 | 0.224 | 0.028 |
| DeepSeek-V4-Flash, Vercel AI Gateway, Upstream provider: Baseten, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.13 | 0.26 | 0.028 |
| DeepSeek-V4-Flash, Vercel AI Gateway, Upstream provider: WaferUpstream provider — the company actually serving this route | 0.13 | 0.23 | 0.125 |
| DeepSeek-V4-Flash, Vercel AI Gateway, Upstream provider: AzureUpstream provider — the company actually serving this route | 0.19 | 0.51 | 0.028 |
| DeepSeek-V4-Flash, Vercel AI Gateway, Upstream provider: GMICloud, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.286 | 0.858 | 0.0091 |
| DeepSeek-V4-Flash, Vercel AI Gateway, Upstream provider: Novita AI, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.44 | 1.32 | 0.028 |
| Deep Infra | 0.09 | 0.18 | 0.018 |
| OpenRouter13 routes · from 0.09 | |||
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: DeepInfra, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.09 | 0.18 | 0.018 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: DigitalOceanUpstream provider — the company actually serving this route | 0.098 | 0.196 | 0.0196 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: SiliconFlow, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.13 | 0.28 | 0.028 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: Crusoe, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.14 | 0.28 | 0.03 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: AtlasCloud, Quantization: FP4Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.14 | 0.28 | 0.028 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: Parasail, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.14 | 0.28 | 0.07 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: Alibaba, Quantization: FP8, Region: us-east-1Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weightsRegion — where this route is served | 0.15 | 0.30 | 0.03 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: Mancer 2, Quantization: FP8Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.19 | 0.50 | Not published |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: Azure, Region: USUpstream provider — the company actually serving this routeRegion — where this route is served | 0.21 | 0.56 | 0.031 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: CloudflareUpstream provider — the company actually serving this route | 0.44 | 1.32 | 0.014 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: Wafer, Quantization: MXFP4Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | 0.60 | 0.17 | 0.20 |
| DeepSeek-V4-Flash, OpenRouter, Upstream provider: Relace, Quantization: FP4Upstream provider — the company actually serving this routeQuantization — numeric precision of the served weights | Not published | Not published | Not published |
| OpenRouter | Not published | Not published | Not published |
| Abacus | 0.14 | 0.28 | 0.03 |
| CoreWeave | 0.14 | 0.28 | 0.07 |
| Nebius Token Factory | 0.14 | 0.28 | Not published |
| Novita | 0.14 | 0.28 | 0.028 |
| Poe | 0.14 | 0.28 | Not published |
| Tensormesh | 0.14 | 0.28 | Free |
| ai& | 0.15 | 0.25 | 0.08 |
| Azure | 0.19 | 0.51 | Not published |
| Qwen Ai Platform | 0.20 | 0.40 | 0.04 |
| InferX | Free | Free | Not published |