Route order log in sign up
1 pick your primary route

DeepSeek: DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash · agent · text+image->text · 1M ctx · added 2026-09-10
$0.0250·0.7514 /M effective hit 90% 2.221T/d

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...

view
openrouter/deepseek/deepseek-v4.1-flash
DeepSeek: DeepSeek V4.1 Flash · agent · text+image->text · 1M ctx
2.221T/d 90% $0.0250·0.7514 – 2mo
sticker
life 429 tok/d dp age
¢ᵉ ¢ⁱ
deepinfra/fp8/flex
$0.1120·0.3360·0.0034 $0.0058·0.3360ⁱ $0.0093·0.3358ᵉ 95% – – 0.3% ✓ 2621ms 85t/s 1d ·
morph/fp8
$0.0220·0.4000·0.0040 $0.0044·0.4000ⁱ $0.0070·0.3993ᵉ 84% – – 2.7% ✓ 1362ms 27t/s 1mo ·
¢ ✓
relace/fp4
$0.0060·0.5000·0.0050 $0.0050·0.5000ⁱ $0.0051·0.4997ᵉ 92% – – 7.3% ✓ 451ms 84t/s 1mo ·
★
deepinfra/fp8
$0.1400·0.4200·0.0042 $0.0072·0.4200ⁱ $0.0151·0.4206ᵉ 92% – – 13% ✓ 676ms 104t/s 2mo ·
♻
streamlake/fp8
$0.1470·0.5880·0.0029 $0.0062·0.5880ⁱ $0.0062·0.5843ᵉ 98% – – 2.9% ◐ 1647ms 78t/s 1d ·
river
$0.1200·0.4800·0.0024 $0.0050·0.4800ⁱ $0.0209·0.4795ᵉ 85% – – 0.6% ✓ 654ms 133t/s ✕12h ·
inference-net/us
$0.1030·0.6000·0.0080 $0.0101·0.6000ⁱ $0.0168·0.5991ᵉ 90% – – 1.6% ✓ 290ms 98t/s 1d ·
wafer/us
$0.0510·0.5670·0.0180 $0.0187·0.5670ⁱ $0.0210·0.5664ᵉ 91% – – 7.3% ✓ 310ms 173t/s 1d ·
inference-net/fp8
$0.1030·0.6000·0.0080 $0.0101·0.6000ⁱ $0.0197·0.5992ᵉ 89% – – 4.0% ✓ 251ms 122t/s 1d ·
coreweave/fp8
$0.1300·0.6500·0.0020 $0.0049·0.6500ⁱ $0.0163·0.6494ᵉ 89% – – 8.0% ✓ 575ms 130t/s 1d ·
atlas-cloud/fp8
$0.1350·0.5400·0.0135 $0.0162·0.5400ⁱ $0.0339·0.5380ᵉ 79% – – 0.4% ◐ 735ms 60t/s 1d ·
cohere/fp8
$0.1950·0.6600·0.0030 $0.0073·0.6600ⁱ $0.0222·0.6591ᵉ 89% – – 4.3% ✓ 306ms 135t/s 1d ·
deepinfra/fp8/priority
$0.2100·0.6300·0.0063 $0.0109·0.6300ⁱ $0.0264·0.6297ᵉ 91% – – 0.7% ✓ 781ms 109t/s 1d ·
open-inference/fp4
$0.0700·0.8000·0.0060 $0.0074·0.8000ⁱ $0.0101·0.7997ᵉ 93% – – 3.4% ✓ 1039ms 109t/s 1d ·
wafer/mxfp4
$0.1900·0.6500·0.0120 $0.0160·0.6500ⁱ $0.0287·0.6493ᵉ 90% – – 3.0% ✓ 314ms 155t/s 1d ·
gmicloud/fp8
$0.1800·0.7200·0.0036 $0.0076·0.7200ⁱ $0.0229·0.7209ᵉ 87% – – 0.1% ◐ 1758ms 108t/s 2mo ·
novita/fp8
$0.1950·0.7800·0.0039 $0.0082·0.7800ⁱ $0.0174·0.7932ᵉ 93% – – 1.3% ✓ 1577ms 97t/s 2mo ·
reka
$0.1200·0.6400·0.0080 $0.0105·0.6400ⁱ $0.0368·0.6369ᵉ 78% – – <0.1% ✓ 465ms 53t/s 1d ·
≈ ◔
alibaba
$0.2100·0.8400·0.0042 $0.0088·0.8400ⁱ $0.0238·0.8393ᵉ 91% – – 5.1% ◐ 755ms 129t/s 1mo ·
fireworks/spot
$0.2100·0.8400·0.0042 $0.0088·0.8400ⁱ $0.0246·0.8393ᵉ 90% – – 2.3% ✓ 642ms 102t/s 1d ·
digitalocean
$0.2250·0.9000·0.0060 $0.0109·0.9000ⁱ $0.0224·0.8995ᵉ 92% – – 4.7% ✓ 641ms 159t/s 1mo ·
◔
deepseek
$0.1500·0.6000·0.0030 · $0.0063·0.6000ⁱ $0.0205·0.9568ᵉ · 93% · – – – · ⚠ 950ms · 110t/s · ✕1mo ·
primeintellect
$0.0390·1.2000·0.0012 $0.0020·1.2000ⁱ $0.0047·1.1996ᵉ 91% – – 3.2% ✓ 944ms 97t/s 1d ·
crusoe/fp8
$0.2900·1.2000·0.0070 $0.0133·1.2000ⁱ $0.0257·1.1995ᵉ 93% – – 4.4% ✓ 593ms 171t/s 1d ·
makora/fp8
$0.2600·1.1700·0.0058 $0.0115·1.1700ⁱ $0.0307·1.1695ᵉ 91% – – 0.6% ✓ 1138ms 120t/s 1d ·
baseten
$0.3000·1.2000·0.0070 $0.0136·1.2000ⁱ $0.0281·1.1997ᵉ 92% – – 2.0% ✓ 576ms 200t/s 1d ·
mancer/fp8
$0.2125·0.7500·0.0500 $0.0536·0.7500ⁱ $0.0817·0.7476ᵉ 79% – – <0.1% ✓ 761ms 80t/s 1d ·
fireworks
$0.3000·1.2000·0.0060 $0.0126·1.2000ⁱ $0.0320·1.1995ᵉ 89% – – 0.4% ✓ 519ms 140t/s 2mo ·
baseten/us
$0.3000·1.2000·0.0070 $0.0136·1.2000ⁱ $0.0337·1.1996ᵉ 94% – – 0.3% ✓ 670ms 197t/s 1d ·
parasail/fp8
$0.3000·1.2000·0.0060 $0.0126·1.2000ⁱ $0.0384·1.1996ᵉ 88% – – 2.1% ✓ 393ms 197t/s 1mo ·
venice/fp8
$0.3000·1.2000·0.0075 $0.0141·1.2000ⁱ $0.0410·1.1996ᵉ 89% – – 0.9% ✓ 471ms 203t/s 1mo ·
phala
$0.2100·0.8400·0.0042 $0.0088·0.8400ⁱ $0.1023·0.8389ᵉ 31% – – <0.1% ✓ 1375ms 110t/s ✕13h ·
modal
$0.3000·1.2000·0.0300 $0.0360·1.2000ⁱ $0.0712·1.1993ᵉ 82% – – <0.1% ✓ 546ms 167t/s 1mo ·
⚡
together
$0.3000·1.2000·0.0060 $0.0126·1.2000ⁱ $0.0763·1.1994ᵉ 82% – – 11% ✓ 249ms 223t/s 1mo ·
siliconflow/fp8
$0.3000·1.2000·0.0060 $0.0126·1.2000ⁱ $0.0949·1.1996ᵉ 71% – – <0.1% ✓ 1649ms 113t/s 1mo ·
io-net/fp8
$0.2850·1.1400·0.0028 · $0.0092·1.1400ⁱ $0.1229·1.1494ᵉ · 58% · – – – · ✓ 2707ms · 5t/s · ✕1mo ·
fireworks/us
$0.4500·1.8000·0.0090 $0.0189·1.8000ⁱ $0.0545·1.7994ᵉ 89% – – 1.0% ✓ 336ms 179t/s 1d ·
≈
baidu/fp8
$0.3000·1.2000·0.0060 $0.0126·1.2000ⁱ $0.1320·1.1970ᵉ 80% – – 0.3% ◐ 1245ms 51t/s 1d ·
parasail/fast
$0.4500·1.8000·0.0090 $0.0189·1.8000ⁱ $0.0793·1.7995ᵉ 85% – – 0.2% ✓ 333ms 134t/s 1d ·
google-vertex/global
$0.3000·1.2000·0.0300 $0.0360·1.2000ⁱ $0.1835·1.1997ᵉ 47% – – 0.1% ✓ 949ms 60t/s 1d ·
reka/fp4
$0.2900·1.1600·0.0290 · $0.0348·1.1600ⁱ – · – · – – – · · 5424ms · 40t/s · ✕1mo ·
baseten/fp8
$0.3000·1.2000·0.0300 · $0.0360·1.2000ⁱ – · – · – – – · · 420ms · 64t/s · ✕1mo ·
makora
$0.3000·1.2000·0.0060 · $0.0126·1.2000ⁱ – · – · – – – · · 1471ms · 34t/s · ✕1mo ·
morph/fp4
$0.3000·1.2000·0.0300 · $0.0360·1.2000ⁱ – · – · – – – · · – · – · ✕2mo ·
wafer
$0.3000·1.2000·0.0060 · $0.0126·1.2000ⁱ – · – · – – – · · 619ms · 65t/s · ✕1mo ·
vercel/deepseek/deepseek-v4.1-flash
DeepSeek V4.1 Flash · agent · text+image+pdf->text · 1M ctx
– 90% $0.0288·0.8900 4.9% 2mo
requesty/deepseek/deepseek-v4.1-flash
deepseek-v4.1-flash · agent · text+image->text · 1M ctx
– 88% $0.0465·1.2000 4.2% 2mo
hf/deepseek-ai/DeepSeek-V4.1-Flash
DeepSeek-V4.1-Flash · agent · text+image->text · 1M ctx
– $0.3000·1.2000 – 2mo
compare all models →