Route order log in sign up
1 pick your primary route

DeepSeek: DeepSeek V4 Pro 0813

deepseek/deepseek-v4-pro-0813 · agent · 1M ctx · added 2026-08-12
$0.2184·3.5081 /M effective hit 77% 110.99B/d

DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.

view
openrouter/deepseek/deepseek-v4-pro-0813
DeepSeek: DeepSeek V4 Pro 0813 · agent · 1M ctx
110.99B/d 77% $0.2184·3.5081 – 3mo
sticker
life 429 tok/d dp age
¢ᵉ
streamlake
$0.6600·1.9800·0.0220 $0.0573·1.9800ⁱ $0.1531·1.9791ᵉ 81% – – 10% ◐ 1484ms 47t/s 2mo ·
♻
phala
$0.9570·2.8776·0.0990 $0.1464·2.8776ⁱ $0.1504·2.9719ᵉ 94% – – 0.7% ✓ 5797ms 34t/s ✕12h ·
novita/fp8
$0.9900·2.9700·0.0330 $0.0859·2.9700ⁱ $0.1802·2.9696ᵉ 87% – – 10% ✓ 1944ms 58t/s 2mo ·
gmicloud/fp8
$1.0560·3.1680·0.0352 $0.0916·3.1680ⁱ $0.1671·3.1677ᵉ 90% – – 1.3% ◐ 2905ms 54t/s 2mo ·
ionstream
$0.9600·2.8800·0.0880 · $0.1362·2.8800ⁱ $0.2158·2.8794ᵉ · 85% · – – – · ✓ 2342ms · 25t/s · ✕1mo ·
¢ⁱ ≈ ◔
deepseek
$0.6600·1.9800·0.0220 $0.0573·1.9800ⁱ $0.2516·2.9607ᵉ 79% – – 13% ⚠ 1494ms 50t/s 3mo ·
wafer/us
$0.1110·4.4400·0.0970 $0.0978·4.4400ⁱ $0.1045·4.4385ᵉ 55% – – 17% ✓ 340ms 38t/s 1d ·
deepinfra/fp8
$1.3000·2.6000·0.1000 $0.1663·2.6000ⁱ $0.3185·2.5995ᵉ 80% – – 1.6% ✓ 1136ms 36t/s 2mo ·
≈ ◔
alibaba
$1.1220·3.3660·0.1122 $0.1680·3.3660ⁱ $0.2446·3.3646ᵉ 74% – – 1.1% ◐ 1223ms 31t/s 2mo ·
⚡ ★
digitalocean
$1.3200·3.9600·0.0440 $0.1145·3.9600ⁱ $0.1880·3.9597ᵉ 88% – – 18% ✓ 1162ms 99t/s 2mo ·
sail-research/fp4
$1.3500·2.4000·0.1950 $0.2588·2.4000ⁱ $0.3739·2.3993ᵉ 82% – – 1.6% ✓ 418ms 46t/s 2mo ·
fireworks
$1.3200·3.9600·0.0440 · $0.1145·3.9600ⁱ $0.2144·3.9596ᵉ · 87% · – – – · ✓ 983ms · 55t/s · ✕1mo ·
sail-research/us
$1.3500·2.4000·0.1950 $0.2588·2.4000ⁱ $0.3962·2.3996ᵉ 72% – – 2.2% ✓ 667ms 55t/s 1mo ·
nebius/fp8
$1.3200·3.9600·0.0400 $0.1108·3.9600ⁱ $0.2505·3.9591ᵉ 86% – – 1.3% ✓ 462ms 61t/s 1d ·
nextbit/fp8
$1.1220·3.3660·0.0370 · $0.0970·3.3660ⁱ $0.3791·3.3657ᵉ · 68% · – – – · ✓ 2099ms · 46t/s · ✕1mo ·
cloudflare
$1.3200·3.9600·0.0440 $0.1145·3.9600ⁱ $0.3148·3.9579ᵉ 81% – – 0.4% ✓ 827ms 30t/s 2mo ·
wafer/mxfp4
$0.2850·4.3000·0.2840 $0.2841·4.3000ⁱ $0.2845·4.2961ᵉ 51% – – 2.0% ✓ 306ms 34t/s 1d ·
coreweave/fp8
$1.3100·3.9600·0.0440 $0.1140·3.9600ⁱ $0.3333·3.9597ᵉ 76% – – 6.7% ✓ 805ms 70t/s 2mo ·
together
$1.3200·3.9600·0.1300 $0.1958·3.9600ⁱ $0.3405·3.9594ᵉ 82% – – 4.9% ✓ 640ms 94t/s 2mo ·
≈
baidu/fp8
$1.3200·3.9600·0.0420 $0.1126·3.9600ⁱ $0.4340·3.9588ᵉ 79% – – 4.4% ◐ 1050ms 30t/s 2mo ·
¢ ✓
relace/fp4
$0.0999·8.0000·0.0999 $0.0999·8.0000ⁱ $0.0998·8.0000ᵉ 58% – – 1.8% ✓ 661ms 18t/s 1d ·
venice
$1.6500·4.9500·0.1650 $0.2471·4.9500ⁱ $0.5076·4.9457ᵉ 65% – – 0.1% ✓ 1312ms 23t/s 1mo ·
parasail/fp8
$1.3200·3.9600·0.0440 $0.1145·3.9600ⁱ $0.6294·3.9596ᵉ 65% – – 0.4% ✓ 481ms 53t/s 2mo ·
≈ ◔
alibaba/us-east-1
$1.3200·3.9600·0.1320 $0.1977·3.9600ⁱ $0.6828·3.9574ᵉ 48% – – 0.2% ◐ 1227ms 21t/s 1d ·
baseten/fp4
$1.3200·3.9600·0.1320 $0.1977·3.9600ⁱ $0.7404·3.9596ᵉ 48% – – 0.2% ✓ 675ms 51t/s 2mo ·
parasail/fast
$1.9800·5.9400·0.0660 $0.1718·5.9400ⁱ $0.6613·5.9380ᵉ 75% – – 0.4% ✓ 469ms 72t/s 1d ·
siliconflow/fp8
$1.3200·3.9600·0.0440 $0.1145·3.9600ⁱ $1.0055·3.9574ᵉ 28% – – 0.1% ✓ 1017ms 17t/s 2mo ·
atlas-cloud/fp8
$1.3200·3.9600·0.1320 $0.1977·3.9600ⁱ $1.1269·3.9541ᵉ 18% – – <0.1% ◐ 1645ms 18t/s 1d ·
ionstream/fp4
$1.1000·3.3000·0.0360 · $0.0948·3.3000ⁱ – · – · – – – · · 870ms · 86t/s · ✕2mo ·
vercel/deepseek/deepseek-v4-pro-0813
DeepSeek V4 Pro 0813 · agent · 1M ctx
– 80% $0.2675·3.9600 – 3mo
requesty/deepseek/deepseek-v4-pro-0813
deepseek-v4-pro-0813 · agent · 1M ctx
– 92% $0.5278·3.5000 0.2% 3mo
hf/deepseek-ai/DeepSeek-V4-Pro-0813
DeepSeek-V4-Pro-0813 · agent · 1M ctx
– $1.3200·3.9600 – 3mo
compare all models →