Every provider of DeepSeek V4.1 Flash from DeepSeek that OpenRouter, Vercel AI Gateway and models.dev list, cheapest first: input, output and cached-input prices per 1M tokens, context, speed and uptime. Checked every hour.
Pareto Inference runs this page and is one of the providers listed. Rows are sorted by price, and Pareto's rows are labeled and placed by price like any other. The methodology says what is included and what is left out. Spot a wrong number? Email us.
Cheapest right now
USD per 1M tokens at 17:32 UTC, peak hours applied.
Agent blend$0.005516Baidu (fp8), OpenRouter
3:1 blend$0.0525engy, Direct
Input$0.006Relace (fp4), OpenRouter
Cached input$0.001014Baidu (fp8), OpenRouter
Output$0.12engy, Direct
Providers
Sorted by the agent blend now: the price of 1M tokens that are 5% uncached input, 94% cached input and 1% output, the mix measured on Pareto's coding-agent traffic. Under it, the 3:1 blend: three input tokens to one output token, cache ignored. Speed and time to first token are Vercel AI Gateway's medians for the last hour; uptime is the last day's, from Vercel AI Gateway where it lists the provider, else OpenRouter.
| Provider | Blended | Input | Output | Cached input | Context | Speed | Uptime |
|---|---|---|---|---|---|---|---|
| Baidufp8OpenRouter | $0.0055163:1 $0.08872 | $0.0507 | $0.2028 | $0.001014 | 1.05M393K out | — | 100% |
| Pareto InferenceVercel AI GatewayPeak: 01:00–04:00 and 06:00–10:00 UTC, Mon–Fri | $0.008163:1 $0.1312 | $0.075$0.15 peak | $0.30$0.60 peak | $0.0015$0.003 peak | 1M384K out | 71 tok/s2.39 s TTFT | 100% |
| ZenMuxDirect, per models.dev | $0.008163:1 $0.1312 | $0.075 | $0.30 | $0.0015 | 1M384K out | — | — |
| engyDirect, per models.dev | $0.008343:1 $0.0525 | $0.03 | $0.12 | $0.006 | 1.05M65.5K out | — | — |
| Morphfp8OpenRouter · Vercel AI Gateway | $0.008863:1 $0.1165 | $0.022 | $0.40 | $0.004 | 1.05M944K out | 110 tok/s1.09 s TTFT | 99.86% |
| NanoGPTDirect, per models.dev | $0.008863:1 $0.1165 | $0.022 | $0.40 | $0.004 | 1.05M384K out | — | — |
| Relacefp4OpenRouter | $0.013:1 $0.1295 | $0.006 | $0.50 | $0.005 | 1.05M944K out | — | 100% |
| OpenInferencefp4OpenRouter | $0.010643:1 $0.1631 | $0.0175 | $0.60 | $0.004 | 1.05M944K out | — | 100% |
| Pareto InferenceDirectPeak: 01:00–04:00 and 06:00–10:00 UTC, Mon–Fri, except Chinese public holidays | $0.011423:1 $0.1837 | $0.105$0.21 peak | $0.42$0.84 peak | $0.0021$0.0042 peak | 1M131K out | — | — |
| RelaceVercel AI Gateway | $0.011453:1 $0.1512 | $0.035 | $0.50 | $0.005 | 1.05M1.05M out | 89 tok/s0.49 s TTFT | 99.93% |
| DeepInfrafp8 flexOpenRouter | $0.012123:1 $0.168 | $0.112 | $0.336 | $0.00336 | 1.05M393K out | — | 29.3% |
| AlibabaOpenRouterPeak: 00:00–14:00 UTC | $0.01383:1 $0.1312 | $0.075$0.15 peak | $0.30$0.60 peak | $0.0075$0.015 peak | 1M393K out | — | 100% |
| KenariDirect, per models.dev | $0.014493:1 $0.1964 | $0.1571 | $0.3143 | $0.003714 | 1M384K out | — | — |
| DevPass (LLM Gateway)Direct, per models.dev | $0.014693:1 $0.2363 | $0.135 | $0.54 | $0.0027 | 1.05M384K out | — | — |
| CoreWeavefp8OpenRouter | $0.014883:1 $0.26 | $0.13 | $0.65 | $0.002 | 1.05M944K out | — | 100% |
| DeepInfrafp8OpenRouter · Vercel AI Gateway | $0.015153:1 $0.21 | $0.14 | $0.42 | $0.0042 | 1.05M393K out | 137 tok/s1.35 s TTFT | 99.96% |
| Standard ComputeDirect, per models.dev | $0.015153:1 $0.21 | $0.14 | $0.42 | $0.0042 | 1M50K out | — | — |
| StreamLakefp8OpenRouter | $0.015993:1 $0.2572 | $0.147 | $0.588 | $0.00294 | 1.02M384K out | — | 100% |
| CommandCodeDirect, per models.dev | $0.016323:1 $0.2625 | $0.15 | $0.60 | $0.003 | 1M384K out | — | — |
| DeepSeekVercel AI GatewayPeak: 01:00–04:00 and 06:00–10:00 UTC, Mon–Fri, except Chinese public holidays | $0.016323:1 $0.2625 | $0.15$0.30 peak | $0.60$1.20 peak | $0.003$0.006 peak | 1M384K out | 132 tok/s1.39 s TTFT | 99.32% |
| Eden AIDirect, per models.dev | $0.016323:1 $0.2625 | $0.15 | $0.60 | $0.003 | 1M384K out | — | — |
| Merge GatewayDirect, per models.dev | $0.016323:1 $0.2625 | $0.15 | $0.60 | $0.003 | 1M384K out | — | — |
| Ollama CloudDirect, per models.dev | $0.016323:1 $0.2625 | $0.15 | $0.60 | $0.003 | 1.05M384K out | — | — |
| Tencent CloudVercel AI GatewayPeak: 01:00–04:00 and 06:00–10:00 UTC, Mon–Fri | $0.016323:1 $0.2625 | $0.15$0.30 peak | $0.60$1.20 peak | $0.003$0.006 peak | 1M384K out | 97 tok/s1.83 s TTFT | 100% |
| AIHubMixDirect, per models.dev | $0.016863:1 $0.2712 | $0.155 | $0.62 | $0.0031 | 1M384K out | — | — |
| above.devDirect, per models.dev | $0.017953:1 $0.2888 | $0.165 | $0.66 | $0.0033 | 1M384K out | — | — |
| InferenceNetus, fp8OpenRouter | $0.018673:1 $0.2273 | $0.103 | $0.60 | $0.008 | 1.04M393K out | — | 100% |
| GMICloudfp8OpenRouter · Vercel AI Gateway | $0.019583:1 $0.315 | $0.18 | $0.72 | $0.0036 | 1.05M944K out | 109 tok/s1.01 s TTFT | 100% |
| IteraComputeDirect, per models.dev | $0.02023:1 $0.3025 | $0.17 | $0.70 | $0.005 | 1.05M393K out | — | — |
| Novitafp8OpenRouter · Vercel AI Gateway | $0.021223:1 $0.3412 | $0.195 | $0.78 | $0.0039 | 1.05M393K out | 106 tok/s1.23 s TTFT | 100% |
| DeepInfrafp8 priorityOpenRouter | $0.022723:1 $0.315 | $0.21 | $0.63 | $0.0063 | 1.05M393K out | — | 99.13% |
| FireworksspotOpenRouter | $0.022853:1 $0.3675 | $0.21 | $0.84 | $0.0042 | 1.05M944K out | — | 99.65% |
| PhalaOpenRouter | $0.022853:1 $0.3675 | $0.21 | $0.84 | $0.0042 | 1.05M944K out | — | 100% |
| RunwareVercel AI Gateway | $0.02293:1 $0.2625 | $0.15 | $0.60 | $0.01 | 1.05M1.05M out | 89 tok/s3.32 s TTFT | 99.93% |
| RekaOpenRouter | $0.023123:1 $0.33 | $0.12 | $0.96 | $0.008 | 1.05M393K out | — | 99.09% |
| ArceeDirect, per models.dev | $0.024183:1 $0.33 | $0.22 | $0.66 | $0.007 | 1M384K out | — | — |
| PrimeIntellectOpenRouter | $0.024313:1 $0.417 | $0.156 | $1.20 | $0.0048 | 1.05M131K out | — | 99.79% |
| AtlasCloudfp8OpenRouter | $0.024843:1 $0.2363 | $0.135 | $0.54 | $0.0135 | 1.05M393K out | — | 100% |
| WaferusOpenRouter | $0.025143:1 $0.18 | $0.051 | $0.567 | $0.018 | 1.05M944K out | — | 100% |
| DigitalOceanOpenRouter | $0.025893:1 $0.3938 | $0.225 | $0.90 | $0.006 | 1.05M944K out | — | 100% |
| VancineDirect, per models.dev | $0.026113:1 $0.42 | $0.24 | $0.96 | $0.0048 | 1M384K out | — | — |
| Wafermxfp4OpenRouter · Vercel AI Gateway | $0.027283:1 $0.305 | $0.19 | $0.65 | $0.012 | 1.05M944K out | 134 tok/s2.07 s TTFT | 100% |
| AlibabaVercel AI GatewayPeak: 00:00–14:00 UTC | $0.02763:1 $0.2625 | $0.15$0.30 peak | $0.60$1.20 peak | $0.015$0.03 peak | 1M384K out | 122 tok/s2.93 s TTFT | 99.47% |
| BoundlessVercel AI Gateway | $0.02943:1 $0.40 | $0.20 | $1.00 | $0.01 | 1M384K out | 82 tok/s0.44 s TTFT | 100% |
| Coherefp8OpenRouter | $0.029823:1 $0.525 | $0.30 | $1.20 | $0.003 | 1.05M384K out | — | 100% |
| Makorafp8OpenRouter | $0.030153:1 $0.4875 | $0.26 | $1.17 | $0.0058 | 1.05M944K out | — | 99.65% |
| InferenceNetVercel AI Gateway | $0.031583:1 $0.475 | $0.30 | $1.00 | $0.007 | 1.04M384K out | 83 tok/s0.41 s TTFT | 100% |
| AI/ML APIDirect, per models.dev | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M384K out | — | — |
| AzureDirect, per models.dev | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1M128K out | — | — |
| CrossModelDirect, per models.dev | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1M384K out | — | — |
| DigitalOceanVercel AI Gateway | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M1.05M out | 81 tok/s1.39 s TTFT | 100% |
| FireworksOpenRouter · Vercel AI Gateway | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M944K out | 164 tok/s0.66 s TTFT | 100% |
| Kilo GatewayDirect, per models.dev | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M944K out | — | — |
| OfoxDirect, per models.dev | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1M384K out | — | — |
| OpenCodeDirect, per models.dev | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1M384K out | — | — |
| Parasailfp8OpenRouter · Vercel AI Gateway | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M944K out | 100 tok/s0.92 s TTFT | 99.96% |
| SiliconFlowfp8OpenRouter | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M393K out | — | 100% |
| TogetherOpenRouter · Vercel AI Gateway | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M944K out | 177 tok/s0.50 s TTFT | 99.98% |
| WallyDirect, per models.dev | $0.032643:1 $0.525 | $0.30 | $1.20 | $0.006 | 1.05M65.5K out | — | — |
| MeliousDirect, per models.dev | $0.032783:1 $0.446 | $0.223 | $1.115 | $0.01115 | 1M384K out | — | — |
| RunInfraVercel AI Gateway | $0.03283:1 $0.20 | $0.112 | $0.464 | $0.024 | 1.05M1.05M out | 130 tok/s0.39 s TTFT | 100% |
| Crusoefp8OpenRouter | $0.033083:1 $0.5175 | $0.29 | $1.20 | $0.007 | 1.05M944K out | — | 100% |
| BaseTenusOpenRouter · Vercel AI Gateway | $0.033583:1 $0.525 | $0.30 | $1.20 | $0.007 | 1.05M262K out | 210 tok/s0.51 s TTFT | 99.99% |
| Venicefp8OpenRouter | $0.034053:1 $0.525 | $0.30 | $1.20 | $0.0075 | 1M262K out | — | 98.89% |
| Nebiusfp8OpenRouter | $0.03643:1 $0.525 | $0.30 | $1.20 | $0.01 | 1.05M944K out | — | 99.62% |
| VivgridDirect, per models.dev | $0.03723:1 $0.54 | $0.31 | $1.23 | $0.01 | 1M384K out | — | — |
| GreenPTDirect, per models.dev | $0.037573:1 $0.5111 | $0.2556 | $1.278 | $0.01278 | 1M384K out | — | — |
| Alibaba (China)Direct, per models.dev | $0.040773:1 $0.5207 | $0.2975 | $1.19 | $0.01488 | 1M384K out | — | — |
| Particle AIVercel AI Gateway | $0.04173:1 $0.3025 | $0.07 | $1.00 | $0.03 | 1.05M1.05M out | 103 tok/s0.46 s TTFT | 100% |
| FireworksusOpenRouter | $0.048963:1 $0.7875 | $0.45 | $1.80 | $0.009 | 1.05M944K out | — | 100% |
| ParasailfastOpenRouter | $0.048963:1 $0.7875 | $0.45 | $1.80 | $0.009 | 1.05M944K out | — | 100% |
| TinfoilDirect, per models.dev | $0.05393:1 $0.8125 | $0.60 | $1.45 | $0.01 | 1.05M384K out | — | — |
| GoogleglobalOpenRouter · Vercel AI Gateway | $0.05523:1 $0.525 | $0.30 | $1.20 | $0.03 | 1.05M944K out | 126 tok/s0.36 s TTFT | 99.92% |
| ModalOpenRouter · Vercel AI Gateway | $0.05523:1 $0.525 | $0.30 | $1.20 | $0.03 | 1.05M944K out | 148 tok/s0.92 s TTFT | 100% |
| NeonDirect, per models.dev | $0.05523:1 $0.525 | $0.30 | $1.20 | $0.03 | 1M384K out | — | — |
| SyntheticDirect, per models.dev | $0.07023:1 $0.75 | $0.60 | $1.20 | $0.03 | 524K65.5K out | — | — |
| RequestyDirect, per models.dev | $0.0873:1 $0.75 | $0.50 | $1.50 | $0.05 | 1.05M393K out | — | — |
| CortecsDirect, per models.dev | $0.09753:1 $0.525 | $0.30 | $1.20 | $0.075 | 1.05M393K out | — | — |
| AKI.IODirect, per models.dev | $0.13073:1 $0.4907 | $0.2804 | $1.121 | $0.1121 | 1.05M81.9K out | — | — |
| Mancer 2fp8OpenRouter | $0.15353:1 $0.2375 | $0.15 | $0.50 | $0.15 | 1.05M944K out | — | 99.89% |
| VultrDirect, per models.dev | $0.15453:1 $0.2625 | $0.15 | $0.60 | — | 1.05M384K out | — | — |
| TensorXDirect, per models.dev | $0.16223:1 $0.75 | $0.50 | $1.50 | $0.13 | 1.05M384K out | — | — |
| Hugging FaceDirect, per models.dev | $0.3093:1 $0.525 | $0.30 | $1.20 | — | 1.05M384K out | — | — |
- OpenRouter: checked
- Vercel AI Gateway: checked
- models.dev: checked
- Pareto: its own self-serve prices, current
Index