Nemotron 3 Ultra (550B A55B)
Nemotron 3 Ultra is NVIDIA's frontier-scale open model with 550B total / 55B active parameters, built for agentic reasoning, long-context analysis, tool use, and high-stakes RAG. It uses a hybrid Latent Mixture-of-Experts (LatentMoE) architecture interleaving Mamba-2, MoE, and select Attention layers, with Multi-Token Prediction (MTP) for native speculative decoding, and is pre-trained on ~20T tokens with an NVFP4 recipe. Reasoning is configurable on/off (plus a medium-effort mode) via the chat template. It supports up to a 1M-token context and 10 languages (English, French, Spanish, Italian, German, Japanese, Hindi, Korean, Brazilian Portuguese, Chinese). Released with open weights, training data, and recipes under the OpenMDW-1.1 license.
Identité
| Organisation | NVIDIA |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2026-06-04 |
| Licence | openmdw license v1.1 |
| Licence commerciale | oui |
| Licence open source | oui |
| Poids ouverts | oui |
| Auto-hébergement (DGX Spark) | non — 550 Md > 107 Md |
| Modalités | image|text |
| Fenêtre de contexte | 1 000 000 |
| KV cache / token (Q8) | — |
| Paramètres totaux | 550 B |
| Paramètres actifs (session) | 55 B |
| Source des paramètres actifs | actifs : nom (heuristique) ; total : nom (heuristique) |
| Curation des paramètres actifs | 2026-11-04 |
| Prix par session | 0.94 $ (1 000 000 tokens entrée + 200 000 sortie, prix le moins cher par fournisseur) |
| Tokens / tâche (Data Players) | 1.2 M |
| Tokens / tâche (Terminal-Bench) | 83.3 M (est.) |
| Ratio Terminal-Bench / Data Players | 69.4× |
| Coût / tâche (Terminal-Bench) | 7.89 $ (est.) |
| Tokens d'entraînement | 20 000 000 000 000 |
| Coupure de connaissance | 2025-09-30 |
| Créé le | 2026-06-16T23:57:38.252458Z |
| Mis à jour le | 2026-11-04T06:04:58.613077Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/nemotron-3-ultra-550b-a55b |
Prix par fournisseur
Statut du modèle (offres + auto-hébergement) : non — hors DGX Spark (> 107 Md).
Toutes sources confondues : catalogue llm-stats, routeurs EUrouter/Cortecs et catalogues directs (OVHcloud, Scaleway, Infomaniak). Les prix sont convertis en USD pour la comparaison ; le prix publié et le taux BCE utilisé sont dans l'infobulle. Les prix inconnus restent indiqués « — ».
| Fournisseur | Source | Souveraineté | Entrée (/M tok) | Sortie (/M tok) | Quantification | Contexte | Latence (s) | Débit (tok/s) | Statut |
|---|---|---|---|---|---|---|---|---|---|
| DeepInfra | llm-stats | non | 0.50 $ | 2.20 $ | — | — | — | — | active |
Souveraineté qualifiée par offre (modèle × fournisseur) ; survolez le badge pour les critères.
SouverainetéScores par catégorie (top_scores)
| Catégorie | Score |
|---|---|
| chat | 0.60 |
| instruction_following | 0.80 |
| knowledge | 0.70 |
| language | 0.80 |
| legal | 0.70 |
| long_context | 0.70 |
| math | 0.60 |
| physics | 0.70 |
| reasoning | 0.60 |
| science | 0.80 |
| search | 0.40 |
| structured_output | 0.60 |
| finance | 0.60 |
| frontend_development | 0.70 |
| general | 0.60 |
| healthcare | 0.80 |
| agents | 0.50 |
| biology | 0.70 |
| chemistry | 0.70 |
| code | 0.70 |
| communication | 0.60 |
| tool_calling | 0.40 |
| vision | 0.40 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| aa-lcr | reasoning | 0.65 | 1.00 | 0.65 | non | oui | 2026-11-04T06:04:58.727938Z |
| apex | reasoning | 0.85 | 1.00 | 0.85 | non | oui | 2026-11-04T06:04:58.682955Z |
| browsecomp | reasoning | 0.44 | 1.00 | 0.44 | non | oui | 2026-11-04T06:04:58.659952Z |
| critpt | reasoning | 0.03 | 1.00 | 0.03 | non | oui | 2026-11-04T06:04:58.707208Z |
| finance-agent | agents | 0.54 | 1.00 | 0.54 | non | oui | 2026-11-04T06:04:58.752654Z |
| finance-agent-v2 | agents | 0.38 | 1.00 | 0.38 | non | non | 2026-11-04T06:04:58.758337Z |
| gdpval | general | 0.47 | 1.00 | 0.47 | non | oui | 2026-08-11T05:53:38.239177Z |
| gpqa | reasoning | 0.87 | 1.00 | 0.87 | non | oui | 2026-11-04T06:04:58.686249Z |
| humanity's-last-exam | reasoning | 0.37 | 1.00 | 0.37 | non | oui | 2026-11-04T06:04:58.704138Z |
| ifbench | instruction_following | 0.82 | 1.00 | 0.82 | non | oui | 2026-11-04T06:04:58.721368Z |
| imo-answerbench | math | 0.92 | 1.00 | 0.92 | non | oui | 2026-11-04T06:04:58.672028Z |
| livecodebench-v6 | reasoning | 0.89 | 1.00 | 0.89 | non | oui | 2026-11-04T06:04:58.664137Z |
| longbench-v2 | long_context | 0.62 | 1.00 | 0.62 | non | oui | 2026-11-04T06:04:58.738365Z |
| mmlu-pro | language | 0.87 | 1.00 | 0.87 | non | oui | 2026-11-04T06:04:58.710156Z |
| mmlu-prox | language | 0.83 | 1.00 | 0.83 | non | oui | 2026-11-04T06:04:58.743378Z |
| multichallenge | chat | 0.64 | 1.00 | 0.64 | non | oui | 2026-11-04T06:04:58.725089Z |
| omniscience | science | 0.79 | 1.00 | 0.79 | non | oui | 2026-11-04T06:04:58.718402Z |
| pinchbench | code | 0.90 | 1.00 | 0.90 | non | oui | 2026-11-04T06:04:58.640481Z |
| profbench | reasoning | 0.56 | 1.00 | 0.56 | non | oui | 2026-11-04T06:04:58.633578Z |
| ruler | long_context | 0.95 | 1.00 | 0.95 | non | oui | 2026-11-04T06:04:58.732964Z |
| scicode | reasoning | 0.45 | 1.00 | 0.45 | non | oui | 2026-11-04T06:04:58.689886Z |
| swe-bench-multilingual | reasoning | 0.68 | 1.00 | 0.68 | non | oui | 2026-11-04T06:04:58.630245Z |
| swe-bench-verified | reasoning | 0.71 | 1.00 | 0.71 | non | oui | 2026-11-04T06:04:58.627058Z |
| tau3-bench | agents | 0.23 | 1.00 | 0.23 | non | oui | 2026-11-04T06:04:58.656438Z |
| terminal-bench-2.1 | reasoning | 0.56 | 1.00 | 0.56 | non | oui | 2026-11-04T06:04:58.623680Z |
| wmt24++ | language | 0.84 | 1.00 | 0.84 | non | oui | 2026-11-04T06:04:58.746214Z |
Classements
| Catégorie | Rang | Score | Note conservatrice | Benchmarks évalués | Méthode | Daté du | Provenance |
|---|---|---|---|---|---|---|---|
| chat | 20 | 0.60 | 24.33 | 1 | trueskill | 2026-11-04T08:08:43.789169Z | rankings |
| knowledge | 1 | 0.70 | 23.69 | 2 | trueskill | 2026-11-04T08:08:45.139626Z | rankings |
| math | 44 | 0.60 | 36.02 | 6 | trueskill | 2026-11-04T08:08:42.442591Z | rankings |
| agents | 118 | — | 12.14 | 8 | trueskill | 2026-10-19 | model_history |
| code | 108 | — | 20.41 | 5 | trueskill | 2026-10-19 | model_history |
Données
Dates de curation (récupération depuis l'API) : models 2026-11-04 · benchmarks 2026-11-04 · scores 2026-11-04 · updates 2026-11-04 · account 2026-11-04 · model_history 2026-11-04 · rankings 2026-11-04