Sarvam-105B
Sarvam-105B is Sarvam AI's flagship open-source Mixture-of-Experts reasoning model built for complex reasoning, coding, and agentic workflows. It uses 128 sparse experts with Multi-head Latent Attention for efficient long-context inference and was pre-trained on 12 trillion tokens spanning code, mathematics, multilingual, and web data.
Identité
| Organisation | Sarvam AI |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2026-03-06 |
| Licence | apache 2.0 |
| Licence commerciale | oui |
| Licence open source | oui |
| Poids ouverts | oui |
| Auto-hébergement (DGX Spark) | oui — ≤ 107 Md (Q8/FP8) |
| Modalités | — |
| Fenêtre de contexte | 131 072 |
| KV cache / token (Q8) | 2.36 Mo/token |
| Paramètres totaux | 105 B |
| Paramètres actifs (session) | 10.3 B |
| Source des paramètres actifs | web : https://huggingface.co/sarvamai/sarvam-105b |
| Curation des paramètres actifs | 2026-11-04 |
| Prix par session | — |
| 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) | — |
| Tokens d'entraînement | 12 000 000 000 000 |
| Coupure de connaissance | — |
| Créé le | 2026-05-07T16:53:39.565723Z |
| Mis à jour le | 2026-11-04T06:05:12.410147Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/sarvam-105b |
Prix par fournisseur
Statut du modèle (offres + auto-hébergement) : partiel — auto-hébergement 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 « — ».
Aucune donnée.
Souveraineté qualifiée par offre (modèle × fournisseur) ; survolez le badge pour les critères.
SouverainetéScores par catégorie (top_scores)
| Catégorie | Score |
|---|---|
| chat | 0.80 |
| instruction_following | 0.80 |
| language | 0.90 |
| legal | 0.90 |
| math | 0.80 |
| physics | 0.80 |
| reasoning | 0.70 |
| search | 0.50 |
| structured_output | 0.80 |
| finance | 0.90 |
| frontend_development | 0.50 |
| general | 0.70 |
| healthcare | 0.90 |
| agents | 0.50 |
| biology | 0.80 |
| chemistry | 0.80 |
| code | 0.50 |
| creativity | 0.70 |
| vision | 0.10 |
| writing | 0.70 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| aime-2025 | math | 0.97 | 1.00 | 0.97 | non | oui | 2026-11-04T06:05:12.439033Z |
| arena-hard-v2 | general | 0.71 | 1.00 | 0.71 | non | oui | 2026-11-04T06:05:12.425037Z |
| beyond-aime | math | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:05:12.448717Z |
| browsecomp | reasoning | 0.49 | 1.00 | 0.49 | non | oui | 2026-11-04T06:05:12.451788Z |
| gpqa | reasoning | 0.79 | 1.00 | 0.79 | non | oui | 2026-11-04T06:05:12.431988Z |
| hmmt-2025 | math | 0.86 | 1.00 | 0.86 | non | oui | 2026-11-04T06:05:12.442249Z |
| hmmt25 | math | 0.86 | 1.00 | 0.86 | non | oui | 2026-11-04T06:05:12.445524Z |
| humanity's-last-exam | reasoning | 0.11 | 1.00 | 0.11 | non | oui | 2026-11-04T06:05:12.457633Z |
| ifeval | chat | 0.85 | 1.00 | 0.85 | non | oui | 2026-11-04T06:05:12.428466Z |
| livecodebench-v6 | reasoning | 0.72 | 1.00 | 0.72 | non | oui | 2026-11-04T06:05:12.416483Z |
| math-500 | math | 0.99 | 1.00 | 0.99 | non | oui | 2026-11-04T06:05:12.413574Z |
| mmlu | general | 0.91 | 1.00 | 0.91 | non | oui | 2026-11-04T06:05:12.419278Z |
| mmlu-pro | language | 0.82 | 1.00 | 0.82 | non | oui | 2026-11-04T06:05:12.422090Z |
| swe-bench-verified | reasoning | 0.45 | 1.00 | 0.45 | non | oui | 2026-11-04T06:05:12.454890Z |
Classements
| Catégorie | Rang | Score | Note conservatrice | Benchmarks évalués | Méthode | Daté du | Provenance |
|---|---|---|---|---|---|---|---|
| agents | 168 | — | 5.26 | 1 | trueskill | 2026-10-20 | model_history |
| code | 271 | — | -2.05 | 1 | trueskill | 2026-10-20 | 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