DeepSeek V4.1 Flash
DeepSeek · Open weights · Released Sep 10, 2026 · deepseek/deepseek-v4.1-flash
Best $0.02 in / $0.12 out via OpenInference · fp4
1.05M contextVisionToolsReasoningStructured outputPrompt cachingOpen weights
Available from 43 providersUSD / MTOK · PRICES 21 MINUTES AGO · SPEED 10 MINUTES AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | UPTIME |
|---|---|---|---|---|---|---|---|---|
| OPOpenInference | $0.02 | $1.00 | $0.015 | 1.05M | fp4 | 8s° | 25° tok/s | 99.3% |
 EngyDIRECT | $0.03 | $0.12 | $0.006 | 1.05M | — | — | — | — |
 Baidu | $0.054 | $0.215 | $0.001 | 1.05M | fp8 | ✕ | ✕ | 99.9% |
 Relace | $0.07 | $0.50 | $0.008 | 1.05M | — | 0.62s° | 51° tok/s | 100.0% |
 CheaperInferenceDIRECTMARKETPLACE | $0.08 | $0.33 | — | — | — | 2.5s | 131 tok/s | — |
 Decart | $0.09 | $0.18 | $0.018 | 1.05M | fp4 | 0.46s° | 68° tok/s | 100.0% |
| SRSail Research | $0.09 | $0.25 | $0.006 | 1.05M | fp4 | ✕ | ✕ | 100.0% |
| SRSail Research | $0.09 | $0.25 | $0.006 | 1.05M | fp4 | — | — | 100.0% |
 Morph | $0.10 | $0.60 | $0.01 | 1.05M | fp8 | 12.5s° | 74° tok/s | 100.0% |
 Wafer | $0.10 | $0.65 | $0.008 | 1.05M | mxfp4 | 0.65s° | 58° tok/s | 100.0% |
 InferenceNet | $0.103 | $0.715 | $0.008 | 1.04M | fp8 | ✕ | ✕ | 100.0% |
| InferenceNet | $0.103 | $0.715 | $0.008 | 1.04M | fp8 | — | — | 100.0% |
 Reka | $0.105 | $0.84 | $0.007 | 1.05M | — | — | — | 97.0% |
 DeepInfraFLEX | $0.112 | $0.336 | $0.003 | 1.05M | fp8 | — | — | 99.1% |
 AtlasCloud | $0.135 | $0.54 | $0.013 | 1.05M | fp8 | 1.1s° | 153° tok/s | 100.0% |
| DeepInfra | $0.14 | $0.42 | $0.004 | 1.05M | fp8 | 0.66s° | 99° tok/s | 100.0% |
 StreamLake | $0.147 | $0.588 | $0.003 | 1.02M | fp8 | 1.6s° | 167° tok/s | 100.0% |
 CoreWeave | $0.15 | $0.65 | $0.002 | 1.05M | fp8 | ✕ | ✕ | 99.9% |
 DigitalOcean | $0.165 | $0.66 | $0.006 | 1.05M | — | 0.55s° | 320° tok/s | 100.0% |
 ModelRun | $0.17 | $0.67 | $0.006 | 1.05M | fp4 | — | — | 99.4% |
 GMICloud | $0.18 | $0.72 | $0.004 | 1.05M | fp8 | 1.1s° | 180° tok/s | 100.0% |
 Novita | $0.195 | $0.78 | $0.004 | 1.05M | fp8 | ✕ | ✕ | 87.6% |
 Mancer 2 | $0.20 | $1.00 | — | 1.05M | fp8 | 2.3s° | 83° tok/s | 79.9% |
| DeepInfraPRIORITY | $0.21 | $0.63 | $0.006 | 1.05M | fp8 | — | — | 2.0% |
 Fireworks | $0.21 | $0.84 | $0.004 | 1.05M | — | 1.2s° | 122° tok/s | 100.0% |
 Phala | $0.21 | $0.84 | $0.004 | 1.05M | — | 1.3s° | 135° tok/s | 99.2% |
| MAMakora | $0.26 | $1.17 | $0.006 | 1.05M | fp8 | ✕ | ✕ | 99.9% |
 Crusoe | $0.29 | $1.20 | $0.007 | 1.05M | fp8 | — | — | 95.4% |
 BaseTen | $0.30 | $1.20 | $0.007 | 1.05M | — | ✕ | ✕ | 100.0% |
| BaseTen | $0.30 | $1.20 | $0.007 | 1.05M | — | — | — | 100.0% |
 Cohere | $0.30 | $1.20 | $0.003 | 1.05M | — | — | — | 99.7% |
| Fireworks | $0.30 | $1.20 | $0.006 | 1.05M | — | — | — | 100.0% |
 Google | $0.30 | $1.20 | $0.03 | 1.05M | — | ✕ | ✕ | 98.7% |
 Modal | $0.30 | $1.20 | $0.03 | 1.05M | — | 0.48s° | 180° tok/s | 100.0% |
 Nebius | $0.30 | $1.20 | $0.01 | 1.05M | fp8 | — | — | 96.9% |
 Parasail | $0.30 | $1.20 | $0.006 | 1.05M | fp8 | 0.53s° | 88° tok/s | 100.0% |
 PrimeIntellect | $0.30 | $1.20 | $0.006 | 1.05M | — | — | — | 100.0% |
 SiliconFlow | $0.30 | $1.20 | $0.006 | 1.05M | fp8 | 2.5s° | 111° tok/s | 100.0% |
 Together | $0.30 | $1.20 | $0.006 | 1.05M | — | 0.31s° | 270° tok/s | 100.0% |
 Venice | $0.30 | $1.20 | $0.007 | 1M | fp8 | ✕ | ✕ | 100.0% |
 Alibaba | $0.33 | $1.32 | $0.033 | 1M | — | 5.4s° | 83° tok/s | 99.9% |
| Fireworks | $0.45 | $1.80 | $0.009 | 1.05M | — | — | — | 100.0% |
| BaseTenPRIORITY | $0.60 | $2.40 | $0.014 | 1.05M | — | — | — | 99.5% |
TTFT/TPS: median (P50) over trailing 504h · chat10k profile (~10K in / 200 out) · measured against the provider API directly; ° = routed via OpenRouter (adds a hop) · ✕ = probed, endpoint returned no measurable stream · methodology v1.1 · measured 10 minutes ago
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Input price history · best listed price, observed
↓ 85% SINCE SEP 20, 2026
$0.135 · SEP 20, 2026$0.02 · TODAY
Pricing variants · USD / MTok in / out
STANDARD
$0.02 / $1.00
per MTok in / out
BATCH
$0.05 / $0.15
async batch
CACHE READ $0.015
Recent changesRSS ↗
Oct 24, 2026endpoint added · mancer-242 → 43
Oct 23, 2026endpoint added · engy42 → 42
Oct 23, 2026endpoint added · cohere41 → 42
Oct 23, 2026endpoint added · modelrun41 → 42
Oct 23, 2026endpoint removed · ionstream41 → 42
Oct 22, 2026endpoint added · reka40 → 41
Oct 21, 2026endpoint added · google38 → 40
Oct 20, 2026endpoint added · relace35 → 38
About
DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on...
Specifications
Context window1,048,576
Max output943,718
Modalities intext, image
Modalities outtext
LicenseOpen weights
ReleasedSep 10, 2026
TokenizerDeepSeek
WeightsHugging Face ↗
Open weights
safetensors
Related
Embed badge
[](https://modelindex.ai/models/deepseek/deepseek-v4.1-flash)Sources
OpenRouter API21 minutes ago
AGGREGATORHugging Face21 minutes ago
COMMUNITYModelIndex benchmarks12 minutes ago
MEASUREDVerified 21 minutes ago · list prices, not negotiated rates.