Route order log in sign up
1 pick your primary route

MoonshotAI: Kimi K3

moonshotai/kimi-k3 · agent · text+image+video->text · 1M ctx · added 2026-07-16
$0.3252·14.5654 /M effective hit 91% 178.43B/d

Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...

view
openrouter/moonshotai/kimi-k3
MoonshotAI: Kimi K3 · agent · text+image+video->text · 1M ctx
178.43B/d 91% $0.3252·14.5654 – 3mo
sticker
life 429 tok/d dp age
¢ᵉ ¢ⁱ
wafer/us
$0.2760·9.6000·0.1300 $0.1374·9.6000ⁱ $0.1463·9.5995ᵉ 90% – – 10% ✓ 892ms 85t/s 1d ·
≈
relace/fp4
$0.1824·11.0055·0.1800 $0.1801·11.0055ⁱ $0.1802·10.2061ᵉ 93% – – 10% ✓ 1319ms 80t/s 2mo ·
decart/mxfp4
$1.9500·9.7500·0.1950 $0.2834·9.7500ⁱ $0.3172·9.8226ᵉ 92% – – 6.3% ✓ 1426ms 71t/s 1d ·
chutes/mxfp4
$1.8500·9.5000·0.1850 $0.2689·9.5000ⁱ $0.5624·9.4997ᵉ 58% – – <0.1% ◐ 2196ms 22t/s 3mo ·
♻ ★
inference-net/us
$0.2340·15.0000·0.1950 $0.1970·15.0000ⁱ $0.1968·15.0000ᵉ 95% – – 10% ✓ 650ms 90t/s 1d ·
inference-net/fp4
$0.2340·15.0000·0.1950 $0.1970·15.0000ⁱ $0.1981·15.0000ᵉ 92% – – 10% ✓ 664ms 91t/s 1mo ·
¢ ✓ ≈
wafer/fast
$0.1530·16.0000·0.1520 $0.1521·16.0000ⁱ $0.1521·16.0000ᵉ 89% – – 7.7% ✓ 973ms 83t/s 1d ·
morph/fp8
$2.2500·12.6000·0.2610 · $0.3612·12.6000ⁱ $0.6621·11.6496ᵉ · 77% · – – – · ✓ 2851ms · 18t/s · ✕1mo ·
wafer
$0.1920·16.0000·0.1820 $0.1825·16.0000ⁱ $0.1836·16.0000ᵉ 86% – – 3.4% ✓ 939ms 84t/s 3mo ·
digitalocean
$2.5500·12.9500·0.2550 $0.3706·12.9500ⁱ $0.5252·12.9497ᵉ 83% – – 0.4% ✓ 1647ms 43t/s 3mo ·
sail-research/fp4
$0.3450·16.5800·0.1725 $0.1812·16.5800ⁱ $0.1910·16.5798ᵉ 90% – – 0.6% ✓ 1348ms 67t/s 3mo ·
≈
sail-research/fp4/us
$0.3450·16.5800·0.1725 $0.1812·16.5800ⁱ $0.1945·16.5797ᵉ 87% – – 1.4% ✓ 1343ms 82t/s 1d ·
moonshotai
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $0.4066·15.0000ᵉ 95% – – 8.5% ✓ 6743ms 23t/s 1d ·
phala
$2.5500·12.7500·0.2550 $0.3706·12.7500ⁱ $0.6676·12.7494ᵉ 85% – – 0.1% ✓ 1991ms 40t/s ✕12h ·
together
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $0.4688·15.0000ᵉ 93% – – 6.8% ✓ 1315ms 64t/s 3mo ·
pareto-alpha/us
$1.0000·11.2500·0.2250 $0.2640·11.2500ⁱ $0.9185·11.2499ᵉ 34% – – 0.3% ✓ 3085ms 57t/s 1d ·
modal/mxfp4
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $0.5600·15.0000ᵉ 90% – – 2.9% ✓ 1026ms 68t/s 3mo ·
deepinfra/mxfp4
$2.8500·14.2500·0.2850 $0.4142·14.2500ⁱ $0.6803·14.2497ᵉ 80% – – 2.0% ✓ 1194ms 35t/s 1d ·
≈
makora/fp4
$1.0000·15.0000·0.5000 $0.5252·15.0000ⁱ $0.5989·15.0000ᵉ 78% – – 1.5% ✓ 1089ms 53t/s 1d ·
amazon-bedrock/us-east-2
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $0.6294·15.0000ᵉ 83% – – 0.2% ✓ 9179ms 15t/s 1d ·
inference-net/fast
$0.4500·18.0000·0.3150 $0.3218·18.0000ⁱ $0.3220·18.0000ᵉ 94% – – 0.3% ✓ 585ms 75t/s ✕22h ·
inference-net/us/fast
$0.4500·18.0000·0.3150 $0.3218·18.0000ⁱ $0.3235·18.0000ᵉ 94% – – 9.1% ✓ 556ms 96t/s 1d ·
io-net/fp4
$2.3000·14.0000·0.2300 $0.3343·14.0000ⁱ $0.8187·14.0000ᵉ 70% – – 1.0% ✓ 2000ms 48t/s 1d ·
parasail/fp4
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $0.8562·15.0000ᵉ 80% – – 1.2% ✓ 732ms 65t/s 2mo ·
alibaba
$3.4500·17.2500·0.3450 $0.5014·17.2500ⁱ $0.6290·17.2497ᵉ 91% – – 3.4% ◐ 2061ms 33t/s 2mo ·
baseten/fp8
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $0.9429·15.0000ᵉ 78% – – 0.5% ✓ 1724ms 93t/s 3mo ·
fireworks
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $0.9769·15.0000ᵉ 80% – – 0.7% ✓ 1437ms 62t/s 3mo ·
fireworks/us
$4.5000·22.5000·0.4500 $0.6540·22.5000ⁱ $0.8709·22.4996ᵉ 91% – – 1.4% ✓ 1168ms 46t/s 2mo ·
⚡
kt
$3.0000·15.0000·0.3000 $0.4360·15.0000ⁱ $1.8077·15.0000ᵉ 44% · – – <0.1% 65.0M ◐ 572ms 114t/s 11h ·
fireworks/fast
$4.5000·22.5000·0.4500 $0.6540·22.5000ⁱ $1.2261·22.4996ᵉ 85% – – 0.5% ✓ 6943ms 72t/s 3mo ·
parasail/fast
$4.5000·22.5000·0.4500 $0.6540·22.5000ⁱ $3.9632·22.4998ᵉ 23% – – <0.1% ✓ 384ms 81t/s 1d ·
morph/fast
$6.0000·22.5000·0.6000 · $0.8720·22.5000ⁱ $4.0705·22.4998ᵉ · 36% · – – – · ✓ 2541ms · 18t/s · ✕1mo ·
inference-net
$2.1000·10.9500·0.2300 · $0.3242·10.9500ⁱ – · – · – – – · · 1587ms · 57t/s · ✕1mo ·
morph/fp4
$2.5000·14.0000·0.2900 · $0.4013·14.0000ⁱ – · – · – – – · · – · – · ✕2mo ·
makora
$2.5500·12.7500·0.2560 · $0.3716·12.7500ⁱ – · – · – – – · · 839ms · 38t/s · ✕1mo ·
deepinfra/bf16
$2.8500·14.2500·0.2850 · $0.4142·14.2500ⁱ – · – · – – – · · 2624ms · 21t/s · ✕1mo ·
moonshotai/mxfp4
$3.0000·15.0000·0.3000 · $0.4360·15.0000ⁱ – · – · – – – · · 5270ms · 26t/s · ✕1mo ·
vercel/moonshotai/kimi-k3
Kimi K3 · agent · text+image+pdf+video->text · 1M ctx
– 89% $0.5834·15.0000 – 3mo
requesty/moonshot/kimi-k3
kimi-k3 · agent · text+image->text · 1M ctx
– $0.6656·15.0000 – 3mo
hf/moonshotai/Kimi-K3
Kimi-K3 · agent · text+image->text · 1M ctx
– $2.9250·14.6250 – 3mo