Route order log in sign up
1 pick your primary route

DeepSeek: DeepSeek V4 Flash 0731

deepseek/deepseek-v4-flash-0731 · agent · 1M ctx · added 2026-07-31
$0.0701·0.4388 /M effective hit 60% 577.52B/d

DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows....

view
openrouter/deepseek/deepseek-v4-flash-0731
577.52B/d 60% $0.0701·0.4388 – 3mo
sticker
life 429 tok/d dp age
¢ᵉ
open-inference/fp8
$0.0300·0.1300·0.0100 · $0.0107·0.1300ⁱ $0.0153·0.1238ᵉ · 75% · – – – · ✓ 1108ms · 24t/s · ✕1mo ·
decart/fp4
$0.0657·0.1314·0.0131 · $0.0151·0.1314ⁱ $0.0233·0.1354ᵉ · 82% · – – – · ✓ 10405ms · 22t/s · ✕2mo ·
relace/fp4
$0.0700·0.1400·0.0140 $0.0160·0.1400ⁱ $0.0298·0.1389ᵉ 73% – – 12% ✓ 706ms 56t/s 3mo ·
★
deepinfra/fp8
$0.0600·0.1800·0.0150 $0.0166·0.1800ⁱ $0.0368·0.1786ᵉ 49% – – 14% ✓ 591ms 54t/s 2mo ·
streamlake/fp8
$0.1320·0.3960·0.0042 $0.0089·0.3960ⁱ $0.0147·0.3956ᵉ 92% – – 8.9% ◐ 2758ms 36t/s 3mo ·
✓
wafer/us
$0.0434·0.3000·0.0350 $0.0353·0.3000ⁱ $0.0402·0.2985ᵉ 34% – – 14% ✓ 430ms 51t/s 1d ·
ambient/fp4
$0.0800·0.1800·0.0160 · $0.0183·0.1800ⁱ $0.0551·0.1788ᵉ · 39% · – – – · ◐ 727ms · 28t/s · ✕2mo ·
makora
$0.0900·0.1950·0.0196 · $0.0222·0.1950ⁱ $0.0608·0.1943ᵉ · 41% · – – – · ✓ 418ms · 69t/s · ✕1mo ·
akashml/fp8
$0.1400·0.2800·0.0200 · $0.0244·0.2800ⁱ $0.0583·0.2788ᵉ · 68% · – – – · ✓ 645ms · 35t/s · ✕2mo ·
digitalocean
$0.1190·0.2380·0.0238 $0.0273·0.2380ⁱ $0.0658·0.2378ᵉ 62% – – 8.5% ✓ 786ms 92t/s 3mo ·
coreweave/fp8
$0.1300·0.2700·0.0500 $0.0529·0.2700ⁱ $0.0755·0.2698ᵉ 64% – – 8.7% ✓ 469ms 88t/s 3mo ·
together
$0.1400·0.2800·0.0300 · $0.0340·0.2800ⁱ $0.0776·0.2790ᵉ · 57% · – – – · ✓ 835ms · 37t/s · ✕1mo ·
inceptron/fp4
$0.1300·0.4950·0.0270 $0.0308·0.4950ⁱ $0.0548·0.4944ᵉ 79% – – 0.7% ✓ 1586ms 35t/s 3mo ·
baseten/fp8
$0.1300·0.2600·0.0280 $0.0317·0.2600ⁱ $0.0835·0.2596ᵉ 45% – – 0.7% ✓ 466ms 103t/s 3mo ·
venice
$0.1300·0.2600·0.0280 $0.0317·0.2600ⁱ $0.0835·0.2597ᵉ 48% – – 2.7% ✓ 883ms 68t/s 3mo ·
≈
reka
$0.1000·0.4200·0.0600 $0.0615·0.4200ⁱ $0.0789·0.3137ᵉ 52% – – 1.5% ✓ 628ms 71t/s 1d ·
baseten/fp32
$0.1300·0.2600·0.0280 $0.0317·0.2600ⁱ $0.0874·0.2595ᵉ 40% – – 0.3% ✓ 369ms 88t/s 1d ·
morph/bf16
$0.1420·0.3996·0.0359 $0.0398·0.3996ⁱ $0.0728·0.3989ᵉ · 65% · – – – · ✓ 1211ms 8t/s ✕1mo ·
sail-research/us
$0.1500·0.3000·0.0420 $0.0459·0.3000ⁱ $0.0861·0.2981ᵉ 59% – – 0.4% ✓ 1078ms 36t/s 1mo ·
sail-research/fp4
$0.1500·0.3000·0.0420 $0.0459·0.3000ⁱ $0.0918·0.2996ᵉ 60% – – <0.1% ✓ 3228ms 23t/s 3mo ·
cohere/fp8
$0.1400·0.2800·0.0700 $0.0726·0.2800ⁱ $0.0974·0.2793ᵉ 59% – – 6.3% ✓ 471ms 95t/s 1d ·
parasail/fp8
$0.1400·0.2800·0.0500 $0.0533·0.2800ⁱ $0.0979·0.2797ᵉ 51% – – 2.0% ✓ 564ms 68t/s 3mo ·
fireworks
$0.2200·0.6600·0.0070 · $0.0148·0.6600ⁱ $0.0648·0.6597ᵉ · 73% · – – – · ✓ 804ms · 66t/s · ✕1mo ·
gmicloud/fp8
$0.2860·0.8580·0.0091 $0.0192·0.8580ⁱ $0.0655·0.8568ᵉ 79% – – 2.3% ◐ 2252ms 58t/s 3mo ·
mancer/fp8
$0.1500·0.5000·0.0750 $0.0777·0.5000ⁱ $0.1052·0.4997ᵉ 58% – – <0.1% ✓ 855ms 58t/s 3mo ·
wafer/fast
$0.1400·0.2300·0.1350 $0.1352·0.2300ⁱ $0.1374·0.2291ᵉ 52% – – 1.5% ✓ 414ms 64t/s 3mo ·
nebius/fp8
$0.1400·0.2800 $0.1400·0.2800ⁱ $0.1398·0.2791ᵉ · 56% · – – – · ✓ 9556ms 21t/s ✕2mo ·
siliconflow/fp8
$0.2200·0.6600·0.0280 $0.0350·0.6600ⁱ $0.0999·0.6598ᵉ 62% – – 0.6% ✓ 1336ms 20t/s 3mo ·
◔
deepseek
$0.2200·0.6600·0.0070 · $0.0148·0.6600ⁱ $0.0635·1.0464ᵉ · 85% · – – – · ⚠ 957ms · 78t/s · ✕2mo ·
cloudflare
$0.4400·1.3200·0.0140 $0.0295·1.3200ⁱ $0.0525·1.3196ᵉ 91% – – 1.1% ✓ 868ms 60t/s 3mo ·
♻ ≈
phala
$0.3080·0.9240·0.0196 $0.0301·0.9240ⁱ $0.0701·1.3015ᵉ 96% – – <0.1% ✓ 1943ms 49t/s ✕23h ·
nextbit/fp8
$0.3520·1.0560·0.0120 · $0.0244·1.0560ⁱ $0.1073·1.0553ᵉ · 72% · – – – · ✓ 1985ms · 8t/s · ✕1mo ·
≈ ◔
alibaba
$0.1760·0.5280·0.0176 $0.0234·0.5280ⁱ $0.1135·1.0536ᵉ 87% – – 1.9% ◐ 1498ms 70t/s 2mo ·
≈ ◔
alibaba/us-east-1
$0.1870·0.5610·0.0187 $0.0248·0.5610ⁱ $0.1232·1.0957ᵉ 80% – – 0.1% ◐ 1828ms 106t/s 1d ·
io-net/fp8
$0.2490·0.6900·0.1240 · $0.1286·0.6900ⁱ $0.1828·0.6012ᵉ · 46% · – – – · ✓ 919ms · 99t/s · ✕2mo ·
≈
baidu/fp8
$0.0598·0.1795·0.0019 $0.0040·0.1795ⁱ $0.1833·1.3030ᵉ 77% – – 4.5% ◐ 1225ms 66t/s 3mo ·
novita/fp8
$0.4092·1.2276·0.0260 $0.0400·1.2276ⁱ $0.2190·1.2276ᵉ 52% – – 7.1% ✓ 2024ms 98t/s 3mo ·
atlas-cloud/fp4
$0.4400·1.3200·0.0280 $0.0430·1.3200ⁱ $0.2506·1.3193ᵉ 51% – – <0.1% ◐ 2251ms 79t/s 3mo ·
chutes/fp8
$0.4400·1.3200·0.0440 $0.0584·1.3200ⁱ $0.3067·1.3190ᵉ 37% – – <0.1% ◐ 1632ms 56t/s 1d ·
⚡
primeintellect/fp4
$0.4400·1.3200·0.0140 $0.0295·1.3200ⁱ $0.3287·1.3198ᵉ 34% – – <0.1% ✓ 1293ms 156t/s 1d ·
¢ ¢ⁱ
open-inference/fp4
$0.0300·0.1000·0.0070 · $0.0078·0.1000ⁱ – · – · – – – · · 2154ms · 8t/s · ✕2mo ·
deepinfra/fp4
$0.0800·0.1800·0.0160 · $0.0183·0.1800ⁱ – · – · – – – · · 813ms · 32t/s · ✕2mo ·
streamlake
$0.1050·0.2100·0.0210 · $0.0241·0.2100ⁱ – · – · – – – · · 1356ms · 52t/s · ✕3mo ·
reka/fp4
$0.1100·0.6600·0.0070 · $0.0108·0.6600ⁱ – · – · – – – · · 648ms · 117t/s · ✕1mo ·
ionstream/fp4
$0.2100·0.4200·0.1000 · $0.1040·0.4200ⁱ – · – · – – – · · 924ms · 30t/s · ✕3mo ·
◔
deepseek/fp8
$0.2200·0.6600·0.0070 · $0.0148·0.6600ⁱ – · – · – – – · · 1035ms · 69t/s · ✕2mo ·
gmicloud/fp4
$0.2860·0.8580·0.0091 · $0.0192·0.8580ⁱ – · – · – – – · · – · – · ✕2mo ·
vercel/deepseek/deepseek-v4-flash-0731
DeepSeek V4 Flash 0731 · agent · 1M ctx
– 71% $0.0631·0.2800 – 3mo
requesty/deepseek/deepseek-v4-flash-0731
deepseek-v4-flash-0731 · agent · 1M ctx
– 68% $0.0451·0.2850 4.8% 3mo
hf/deepseek-ai/DeepSeek-V4-Flash-0731
DeepSeek-V4-Flash-0731 · agent · 1M ctx
– $0.1300·0.2600 – 3mo