Route order log in sign up
1 pick your primary route

NVIDIA: Nemotron 3 Ultra

nvidia/nemotron-3-ultra-550b-a55b · agent · 262k ctx · added 2026-06-04
$0.2462·2.3372 /M effective hit 69% 4.74B/d

NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...

view
openrouter/nvidia/nemotron-3-ultra-550b-a55b
NVIDIA: Nemotron 3 Ultra · agent · 262k ctx
4.74B/d 69% $0.2462·2.3372 – 5mo
sticker
life 429 tok/d dp age
¢ ¢ᵉ ¢ⁱ
deepinfra/fp4/flex
$0.4000·1.7600·0.0800 $0.0980·1.7600ⁱ $0.1281·1.7567ᵉ 79% – – 0.1% ✓ – – 1d ·
✓ ★
deepinfra/fp4
$0.5000·2.2000·0.1000 $0.1225·2.2000ⁱ $0.2068·2.1992ᵉ 68% – – 39% ✓ 926ms 59t/s 3mo ·
baseten/fp4/us
$0.6000·2.4000·0.1200 $0.1470·2.4000ⁱ $0.2678·2.3989ᵉ 75% – – 26% ✓ 280ms 126t/s 1d ·
♻
venice/fp8
$0.6250·3.1250·0.1875 $0.2121·3.1250ⁱ $0.2361·3.1243ᵉ 81% – – 18% ✓ 1136ms 67t/s 3mo ·
baseten/fp4
$0.6000·2.4000·0.1200 $0.1470·2.4000ⁱ $0.3705·2.3984ᵉ 57% – – 13% ✓ 220ms 110t/s 3mo ·
⚡
kt
$0.5000·2.2000·0.1000 $0.1225·2.2000ⁱ $0.3941·2.1997ᵉ 26% – – 3.6% ◐ 962ms 140t/s 13h ·
together
$0.6000·3.6000·0.2000 · $0.2225·3.6000ⁱ $0.3989·3.5998ᵉ · 50% · – – – · ✓ 1569ms · 64t/s · ✕2mo ·
vercel/nvidia/nemotron-3-ultra-550b-a55b
Nemotron 3 Ultra · agent · 1M ctx
– 66% $0.2692·2.4000 – 5mo
requesty/nvidia/nemotron-3-ultra-550b-a55b
nemotron-3-ultra-550b-a55b · agent · 1M ctx
– 76% $0.5000·1.5000 2.5% 5mo
hf/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16
– $1.0000·5.0000 – 5mo