Gemma 4 12B
Gemma 4 12B is Google DeepMind's encoder-free multimodal instruction-tuned model with 11.95 billion parameters and a 256K context window. It supports text, image, audio, and video inputs with text output, projecting image patches and audio waveforms directly into a single decoder-only transformer for streamlined local deployment.
Identité
| Organisation | |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2026-05-23 |
| 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 | 262 144 |
| KV cache / token (Q8) | 196.6 ko/token |
| Paramètres totaux | 12 B |
| Paramètres actifs (session) | 12 B |
| Source des paramètres actifs | actifs = total (modèle dense, config HF (google/gemma-4-12B)) ; total : config Hugging Face (google/gemma-4-12B) |
| 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 | — |
| Coupure de connaissance | 2025-01-01 |
| Créé le | 2026-06-09T17:39:52.736639Z |
| Mis à jour le | 2026-11-04T06:04:46.550159Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/gemma-4-12b-it |
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 |
|---|---|
| language | 0.70 |
| legal | 0.80 |
| long_context | 0.40 |
| math | 0.60 |
| multimodal | 0.50 |
| physics | 0.80 |
| reasoning | 0.60 |
| speech_to_text | 0.70 |
| structured_output | 0.20 |
| finance | 0.80 |
| general | 0.60 |
| healthcare | 0.60 |
| audio | 0.40 |
| biology | 0.80 |
| chemistry | 0.80 |
| vision | 0.40 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| aime-2026 | math | 0.78 | 1.00 | 0.78 | non | oui | 2026-11-04T06:04:46.558965Z |
| big-bench-extra-hard | reasoning | 0.53 | 1.00 | 0.53 | non | oui | 2026-11-04T06:04:46.582965Z |
| codeforces | math | 0.55 | 3 000.00 | 0.55 | non | oui | 2026-11-04T06:04:46.567413Z |
| covost2 | audio | 0.39 | 1.00 | 0.39 | non | oui | 2026-11-04T06:04:46.633212Z |
| fleurs | language | 0.93 | 1.00 | 0.93 | non | oui | 2026-11-04T06:04:46.643201Z |
| gpqa | reasoning | 0.79 | 1.00 | 0.79 | non | oui | 2026-11-04T06:04:46.570810Z |
| humanity's-last-exam | reasoning | 0.05 | 1.00 | 0.05 | non | oui | 2026-11-04T06:04:46.578230Z |
| livecodebench-v6 | reasoning | 0.72 | 1.00 | 0.72 | non | oui | 2026-11-04T06:04:46.563512Z |
| mathvision | math | 0.80 | 1.00 | 0.80 | non | oui | 2026-11-04T06:04:46.617348Z |
| medxpertqa | healthcare | 0.49 | 1.00 | 0.49 | non | oui | 2026-11-04T06:04:46.623192Z |
| mmlu-pro | language | 0.77 | 1.00 | 0.77 | non | oui | 2026-11-04T06:04:46.554589Z |
| mmmlu | language | 0.83 | 1.00 | 0.83 | non | oui | 2026-11-04T06:04:46.593154Z |
| mmmu-pro | vision | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:46.603181Z |
| mrcr-v2-(8-needle) | long_context | 0.43 | 1.00 | 0.43 | non | oui | 2026-11-04T06:04:46.653645Z |
| omnidocbench-1.5 | vision | 0.16 | 1.00 | 0.16 | non | oui | 2026-11-04T06:04:46.613182Z |
Classements
Aucune donnée.
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