Phi-3.5-mini-instruct
Phi-3.5-mini-instruct is a 3.8B-parameter model that supports up to 128K context tokens, with improved multilingual capabilities across over 20 languages. It underwent additional training and safety post-training to enhance instruction-following, reasoning, math, and code generation. Ideal for environments with memory or latency constraints, it uses an MIT license.
Identité
| Organisation | Microsoft |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2024-08-23 |
| Licence | mit |
| Licence commerciale | oui |
| Licence open source | oui |
| Poids ouverts | oui |
| Auto-hébergement (DGX Spark) | oui — ≤ 107 Md (Q8/FP8) |
| Modalités | text |
| Fenêtre de contexte | 131 072 |
| KV cache / token (Q8) | 196.6 ko/token |
| Paramètres totaux | 3.8 B |
| Paramètres actifs (session) | 3.82 B |
| Source des paramètres actifs | actifs = total (modèle dense, config HF (microsoft/Phi-3.5-mini-instruct)) ; total : config Hugging Face (microsoft/Phi-3.5-mini-instruct) |
| 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 | 3 400 000 000 000 |
| Coupure de connaissance | — |
| Créé le | 2026-05-07T16:53:31.229358Z |
| Mis à jour le | 2026-11-04T06:04:53.153229Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/phi-3.5-mini-instruct |
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 | 0.40 |
| language | 0.60 |
| legal | 0.60 |
| long_context | 0.50 |
| math | 0.60 |
| physics | 0.60 |
| psychology | 0.70 |
| reasoning | 0.70 |
| summarization | 0.20 |
| finance | 0.60 |
| general | 0.60 |
| healthcare | 0.60 |
| biology | 0.30 |
| chemistry | 0.30 |
| code | 0.60 |
| creativity | 0.60 |
| writing | 0.40 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| arc-c | reasoning | 0.85 | 1.00 | 0.85 | non | oui | 2026-11-04T06:04:53.163567Z |
| arena-hard | general | 0.37 | 1.00 | 0.37 | non | oui | 2026-11-04T06:04:53.173060Z |
| big-bench-hard | reasoning | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:53.179405Z |
| boolq | language | 0.78 | 1.00 | 0.78 | non | oui | 2026-11-04T06:04:53.190315Z |
| govreport | summarization | 0.26 | 1.00 | 0.26 | non | oui | 2026-11-04T06:04:53.200293Z |
| gpqa | reasoning | 0.30 | 1.00 | 0.30 | non | oui | 2026-11-04T06:04:53.213105Z |
| gsm8k | math | 0.86 | 1.00 | 0.86 | non | oui | 2026-11-04T06:04:53.228205Z |
| hellaswag | reasoning | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:53.233294Z |
| humaneval | reasoning | 0.63 | 1.00 | 0.63 | non | oui | 2026-11-04T06:04:53.237420Z |
| math | math | 0.48 | 1.00 | 0.48 | non | oui | 2026-11-04T06:04:53.243045Z |
| mbpp | reasoning | 0.70 | 100.00 | 0.70 | non | oui | 2026-11-04T06:04:53.253029Z |
| mega-mlqa | language | 0.62 | 1.00 | 0.62 | non | oui | 2026-11-04T06:04:53.267841Z |
| mega-tydi-qa | language | 0.62 | 1.00 | 0.62 | non | oui | 2026-11-04T06:04:53.283211Z |
| mega-udpos | language | 0.47 | 1.00 | 0.47 | non | oui | 2026-11-04T06:04:53.293385Z |
| mega-xcopa | reasoning | 0.63 | 1.00 | 0.63 | non | oui | 2026-11-04T06:04:53.303052Z |
| mega-xstorycloze | reasoning | 0.73 | 1.00 | 0.73 | non | oui | 2026-11-04T06:04:53.313046Z |
| mgsm | math | 0.48 | 1.00 | 0.48 | non | oui | 2026-11-04T06:04:53.323045Z |
| mmlu | general | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:53.329429Z |
| mmlu-pro | language | 0.47 | 1.00 | 0.47 | non | oui | 2026-11-04T06:04:53.332950Z |
| mmmlu | language | 0.55 | 1.00 | 0.55 | non | oui | 2026-11-04T06:04:53.339406Z |
| openbookqa | reasoning | 0.79 | 1.00 | 0.79 | non | oui | 2026-11-04T06:04:53.343002Z |
| piqa | reasoning | 0.81 | 1.00 | 0.81 | non | oui | 2026-11-04T06:04:53.349347Z |
| qasper | reasoning | 0.42 | 1.00 | 0.42 | non | oui | 2026-11-04T06:04:53.353190Z |
| qmsum | summarization | 0.21 | 1.00 | 0.21 | non | oui | 2026-11-04T06:04:53.363009Z |
| repoqa | long_context | 0.77 | 1.00 | 0.77 | non | oui | 2026-11-04T06:04:53.373090Z |
| ruler | long_context | 0.84 | 1.00 | 0.84 | non | oui | 2026-11-04T06:04:53.383020Z |
| social-iqa | reasoning | 0.75 | 1.00 | 0.75 | non | oui | 2026-11-04T06:04:53.393174Z |
| squality | summarization | 0.24 | 1.00 | 0.24 | non | oui | 2026-11-04T06:04:53.403195Z |
| summscreenfd | summarization | 0.16 | 1.00 | 0.16 | non | oui | 2026-11-04T06:04:53.413758Z |
| truthfulqa | general | 0.64 | 1.00 | 0.64 | non | oui | 2026-11-04T06:04:53.423136Z |
| winogrande | reasoning | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:53.429873Z |
Classements
| Catégorie | Rang | Score | Note conservatrice | Benchmarks évalués | Méthode | Daté du | Provenance |
|---|---|---|---|---|---|---|---|
| code | 276 | — | -6.87 | 1 | 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