Qwen2.5 VL 32B Instruct
Qwen2.5-VL is a vision-language model from the Qwen family. Key enhancements include visual understanding (objects, text, charts, layouts), visual agent capabilities (tool use, computer/phone control), long video comprehension with event pinpointing, visual localization (bounding boxes/points), and structured output generation.
Identité
| Organisation | Alibaba Cloud / Qwen Team |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2025-02-28 |
| 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 | 128 000 |
| KV cache / token (Q8) | 131.1 ko/token |
| Paramètres totaux | 33.5 B |
| Paramètres actifs (session) | 33.5 B |
| Source des paramètres actifs | actifs = total (modèle dense, config HF (Qwen/Qwen2.5-VL-32B-Instruct)) ; total : config Hugging Face (Qwen/Qwen2.5-VL-32B-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 | — |
| Coupure de connaissance | — |
| Créé le | 2026-05-07T16:53:35.769215Z |
| Mis à jour le | 2026-11-04T06:05:05.710927Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/qwen2.5-vl-32b |
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 |
|---|---|
| image_to_text | 0.70 |
| language | 0.70 |
| legal | 0.70 |
| long_context | 0.50 |
| math | 0.70 |
| multimodal | 0.60 |
| physics | 0.50 |
| reasoning | 0.60 |
| spatial_reasoning | 0.60 |
| structured_output | 0.80 |
| finance | 0.70 |
| general | 0.60 |
| grounding | 0.60 |
| healthcare | 0.70 |
| agents | 0.40 |
| biology | 0.50 |
| chemistry | 0.50 |
| code | 0.90 |
| text-to-image | 0.80 |
| video | 0.50 |
| vision | 0.60 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| aitz-em | multimodal | 0.83 | 1.00 | 0.83 | non | oui | 2026-11-04T06:05:05.716763Z |
| android-control-high-em | multimodal | 0.70 | 1.00 | 0.70 | non | oui | 2026-11-04T06:05:05.721325Z |
| android-control-low-em | multimodal | 0.93 | 1.00 | 0.93 | non | oui | 2026-11-04T06:05:05.724905Z |
| androidworld-sr | general | 0.22 | 1.00 | 0.22 | non | oui | 2026-11-04T06:05:05.728058Z |
| cc-ocr | vision | 0.77 | 1.00 | 0.77 | non | oui | 2026-11-04T06:05:05.731590Z |
| charadessta | video | 0.54 | 1.00 | 0.54 | non | oui | 2026-11-04T06:05:05.735445Z |
| docvqa | vision | 0.95 | 1.00 | 0.95 | non | oui | 2026-11-04T06:05:05.739506Z |
| gpqa | reasoning | 0.46 | 1.00 | 0.46 | non | oui | 2026-11-04T06:05:05.742972Z |
| humaneval | reasoning | 0.92 | 1.00 | 0.92 | non | oui | 2026-11-04T06:05:05.746596Z |
| infovqa | vision | 0.83 | 1.00 | 0.83 | non | oui | 2026-11-04T06:05:05.749850Z |
| lvbench | vision | 0.49 | 1.00 | 0.49 | non | oui | 2026-11-04T06:05:05.753108Z |
| math | math | 0.82 | 1.00 | 0.82 | non | oui | 2026-11-04T06:05:05.756476Z |
| mathvision | math | 0.38 | 1.00 | 0.38 | non | oui | 2026-11-04T06:05:05.759776Z |
| mathvista-mini | math | 0.75 | 1.00 | 0.75 | non | oui | 2026-11-04T06:05:05.763214Z |
| mbpp | reasoning | 0.84 | 100.00 | 0.84 | non | oui | 2026-11-04T06:05:05.766445Z |
| mmbench-video | video | 0.02 | 1.00 | 0.02 | non | oui | 2026-11-04T06:05:05.770788Z |
| mmlu | general | 0.78 | 1.00 | 0.78 | non | oui | 2026-11-04T06:05:05.774048Z |
| mmlu-pro | language | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:05:05.777286Z |
| mmmu | multimodal | 0.70 | 1.00 | 0.70 | non | oui | 2026-11-04T06:05:05.780676Z |
| mmmu-pro | vision | 0.49 | 1.00 | 0.49 | non | oui | 2026-11-04T06:05:05.784940Z |
| mmstar | vision | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:05:05.788748Z |
| ocrbench-v2-(en) | vision | 0.57 | 1.00 | 0.57 | non | oui | 2026-11-04T06:05:05.792260Z |
| ocrbench-v2-(zh) | vision | 0.59 | 1.00 | 0.59 | non | oui | 2026-11-04T06:05:05.796852Z |
| osworld | multimodal | 0.06 | 1.00 | 0.06 | non | oui | 2026-11-04T06:05:05.800239Z |
| screenspot | vision | 0.89 | 1.00 | 0.89 | non | oui | 2026-11-04T06:05:05.804485Z |
| screenspot-pro | vision | 0.39 | 1.00 | 0.39 | non | oui | 2026-11-04T06:05:05.808183Z |
| videomme-w-o-sub. | multimodal | 0.70 | 1.00 | 0.70 | non | oui | 2026-11-04T06:05:05.812045Z |
| videomme-w-sub. | vision | 0.78 | 1.00 | 0.78 | non | oui | 2026-11-04T06:05:05.815672Z |
Classements
| Catégorie | Rang | Score | Note conservatrice | Benchmarks évalués | Méthode | Daté du | Provenance |
|---|---|---|---|---|---|---|---|
| agents | 194 | — | -3.21 | 3 | trueskill | 2026-10-19 | model_history |
| code | 158 | — | 12.35 | 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