Route order log in sign up
1 pick your primary route

DeepSeek: DeepSeek V4 Flash 0423

deepseek/deepseek-v4-flash · agent · 1M ctx · added 2026-04-24
$0.0508·0.2511 /M effective hit 68% 319.85B/d

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...

view
openrouter/deepseek/deepseek-v4-flash
319.85B/d 68% $0.0508·0.2511 – 6mo
sticker
life 429 tok/d dp age
¢ ¢ᵉ ¢ⁱ
open-inference/fp8
$0.0500·0.1400·0.0140 · $0.0203·0.1400ⁱ $0.0203·0.1387ᵉ · 82% · – – – · ✓ 1154ms · 20t/s · ✕1mo ·
gmicloud/fp8
$0.0910·0.1820·0.0182 $0.0309·0.1820ⁱ $0.0346·0.1824ᵉ 74% – – 7.5% ◐ 1424ms 74t/s 3mo ·
≈
streamlake/fp8
$0.0826·0.1652·0.0165 $0.0281·0.1652ⁱ $0.0406·0.1640ᵉ 67% – – 15% ◐ 3933ms 25t/s 3mo ·
★
deepinfra/fp8
$0.0900·0.1800·0.0180 $0.0306·0.1800ⁱ $0.0412·0.1787ᵉ 66% – – 19% ✓ 751ms 58t/s 2mo ·
sail-research/fp4
$0.0900·0.1800·0.0200 · $0.0322·0.1800ⁱ $0.0438·0.1789ᵉ · 66% · – – – · ✓ 746ms · 25t/s · ✕2mo ·
venice
$0.0966·0.1925·0.0196 $0.0330·0.1925ⁱ $0.0462·0.1914ᵉ 69% – – 7.3% ✓ 986ms 55t/s 3mo ·
wafer/mxfp4
$0.1200·0.1700·0.0240 $0.0408·0.1700ⁱ $0.0552·0.1682ᵉ 71% – – 14% ✓ 352ms 75t/s 1d ·
♻
siliconflow/fp8
$0.1300·0.2800·0.0280 $0.0458·0.2800ⁱ $0.0478·0.2792ᵉ 83% – – 3.0% ✓ 1067ms 41t/s 3mo ·
alibaba/fp8
$0.1340·0.2680·0.0268 · $0.0455·0.2680ⁱ $0.0570·0.2674ᵉ · 72% · – – – · ◐ 850ms · 85t/s · ✕1mo ·
fireworks
$0.1050·0.2100·0.0210 $0.0357·0.2100ⁱ $0.0689·0.2094ᵉ 47% – – 0.5% ✓ 448ms 102t/s 3mo ·
digitalocean
$0.0980·0.1960·0.0196 · $0.0333·0.1960ⁱ $0.0774·0.1943ᵉ · 26% · – – – · ✓ 1226ms · 10t/s · ✕1mo ·
⚡ ✓ ≈
relace/fp4
$0.0500·0.4500·0.0500 $0.0500·0.4500ⁱ $0.0499·0.4495ᵉ 71% – – 17% ✓ 524ms 116t/s 1d ·
novita/fp8
$0.1400·0.2800·0.0280 · $0.0475·0.2800ⁱ $0.0710·0.2785ᵉ · 62% · – – – · ✓ 1498ms · 46t/s · ✕1mo ·
atlas-cloud/fp4
$0.1400·0.2800·0.0280 $0.0475·0.2800ⁱ $0.0826·0.2799ᵉ 55% – – 0.3% ◐ 4323ms 32t/s 3mo ·
nextbit/fp8
$0.1500·0.3500·0.0350 · $0.0551·0.3500ⁱ $0.0790·0.3492ᵉ · 62% · – – – · ✓ 1387ms · 61t/s · ✕1mo ·
≈
baidu/fp8
$0.1400·0.2800·0.0280 $0.0475·0.2800ⁱ $0.0998·0.2792ᵉ 60% – – 5.4% ◐ 993ms 106t/s 3mo ·
parasail/fp8
$0.1400·0.2800·0.0700 $0.0822·0.2800ⁱ $0.1005·0.2791ᵉ 56% – – 8.6% ✓ 663ms 41t/s 3mo ·
azure/us
$0.2100·0.5600·0.0310 $0.0622·0.5600ⁱ $0.0755·0.5588ᵉ 77% – – 1.2% ✓ 1131ms 100t/s 2mo ·
coreweave/fp8
$0.1400·0.2800·0.0700 · $0.0822·0.2800ⁱ $0.1122·0.2782ᵉ · 40% · – – – · ✓ 646ms · 40t/s · ✕2mo ·
mancer/fp8
$0.1500·0.5000·0.0750 $0.0881·0.5000ⁱ $0.0970·0.4987ᵉ 68% – – <0.1% ✓ 1054ms 37t/s 2mo ·
phala
$0.2000·0.4000·0.0700 · $0.0927·0.4000ⁱ $0.1380·0.3994ᵉ · 48% · – – – · ✓ 1565ms · 63t/s · ✕1mo ·
◔
deepseek
$0.2200·0.6600·0.0070 · $0.0442·0.6600ⁱ $0.0731·0.9919ᵉ · 80% · – – – · ⚠ 750ms · 72t/s · ✕2mo ·
cloudflare
$0.4400·1.3200·0.0140 · $0.0884·1.3200ⁱ $0.2300·1.3187ᵉ · 49% · – – – · ◐ 765ms · 63t/s · ✕2mo ·
deepinfra/fp4
$0.0900·0.1800·0.0180 · $0.0306·0.1800ⁱ – · – · – – – · · 621ms · 40t/s · ✕2mo ·
wafer
$0.1000·0.2500·0.0500 · $0.0587·0.2500ⁱ – · – · – – – · · 517ms · 72t/s · ✕1mo ·
ambient/fp4
$0.1400·0.2800·0.0280 · $0.0475·0.2800ⁱ – · – · – – – · · – · – · ✕2mo ·
mancer/fp4
$0.1500·0.5000 · $0.1500·0.5000ⁱ – · – · – – – · · 1483ms · 26t/s · ✕2mo ·
vercel/deepseek/deepseek-v4-flash
DeepSeek V4 Flash · agent · 1M ctx
– 68% $0.0783·0.5850 – 6mo
requesty/deepseek/deepseek-v4-flash
deepseek-v4-flash · agent · 1M ctx
– 88% $0.0983·0.2520 · 1.4% 6mo
hf/deepseek-ai/DeepSeek-V4-Flash
DeepSeek-V4-Flash · agent · 1M ctx
– $0.0900·0.1800 – 6mo