LFM3-VL-2B
LFM3-VL-2B is Liquid AI's open-weight 1.92B vision-language model (1.73B LFM3 language model + new 164M in-house conv/attention vision encoder + 25M projector), released Oct 27, 2026 as part of the LFM3 family. Inputs: text, images, and video (first LFM vision model with video, sampled at 2 fps) at native resolution; text output. Reasons step by step before answering (thinking), with spatial/embodied reasoning and grounding aimed at robotics, plus document/OCR/chart and GUI understanding. 128K context; 16 languages. Vendor reports a 71.4 average across 62 vision benchmarks (vs 57.6 for LFM2.5-VL-3B). Not recommended for coding, multi-turn conversations, or complex reasoning per the model card. Hugging Face: LiquidAI/LFM3-VL-2B (+ GGUF, MLX, ONNX). License: LFM Open License v1.0. Catalog benches are the thinking-mode column from the model card table. No catalog id (recorded here only): SEED-Bench image 78.6, MM-Vet v1 57.4, PerceptionBench 37.1 (catalog id is Moonshot's internal benchmark), HR-Bench 8K 76.2, ZoomBench 62.6, V*Bench 86.4, CV-Bench 88.8, LogicVista 60.4, RefCOCO-M 95.2, Point-Bench 73.8, MindCube tiny 56.5, SAT 80.2, Where2Place 44.9, OpenEQA 76.3, VSI-Bench 63.1, ScreenSpot-v2 desktop/mobile/web 94.6/92.2/89.9, MMMB 85.2, Multilingual MMBench 83.6, MM-ToolSandbox 42.5. No gateway id or price at time of adding, so no provider row. Sources: https://huggingface.co/LiquidAI/LFM3-VL-2B · https://www.liquid.ai/blog/lfm3.
Identité
| Organisation | Liquid AI |
|---|---|
| Pays | — |
| Famille | — |
| Type | llm |
| Date de sortie | 2026-10-27 |
| Licence | lfm open license v1.0 |
| Licence commerciale | oui |
| Licence open source | non |
| Poids ouverts | oui |
| Auto-hébergement (DGX Spark) | oui — ≤ 107 Md (Q8/FP8) |
| Modalités | — |
| Fenêtre de contexte | 128 000 |
| KV cache / token (Q8) | 24.6 ko/token |
| Paramètres totaux | 1.92 B |
| Paramètres actifs (session) | 1.92 B |
| Source des paramètres actifs | actifs = total (modèle dense, config HF (LiquidAI/LFM3-VL-2B)) ; total : config Hugging Face (LiquidAI/LFM3-VL-2B) |
| 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 | 22 000 000 000 000 |
| Coupure de connaissance | — |
| Créé le | 2026-11-02T18:10:22.835785Z |
| Mis à jour le | 2026-11-04T06:04:49.783293Z |
| Source | LLM Stats |
| URL | https://llm-stats.com/models/lfm3-vl-2b |
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 |
| long_context | 0.40 |
| math | 0.70 |
| multimodal | 0.70 |
| reasoning | 0.70 |
| safety | 0.90 |
| spatial_reasoning | 0.70 |
| structured_output | 0.80 |
| embodied | 0.80 |
| general | 0.70 |
| grounding | 0.70 |
| healthcare | 0.50 |
| 3d | 0.70 |
| video | 0.60 |
| vision | 0.70 |
Scores par benchmark
| Benchmark | Catégorie | Score | Max | Normalisé (API) | Vérifié | Auto-déclaré | Daté du |
|---|---|---|---|---|---|---|---|
| activitynet | vision | 0.71 | 1.00 | 0.71 | non | oui | 2026-11-04T06:04:50.043344Z |
| blink | vision | 0.71 | 1.00 | 0.71 | non | oui | 2026-11-04T06:04:49.858595Z |
| chartqa | reasoning | 0.93 | 1.00 | 0.93 | non | oui | 2026-11-04T06:04:49.910400Z |
| charxiv-d | reasoning | 0.88 | 1.00 | 0.88 | non | oui | 2026-11-04T06:04:49.933352Z |
| charxiv-r | reasoning | 0.70 | 1.00 | 0.70 | non | oui | 2026-11-04T06:04:49.944116Z |
| countbench | vision | 0.93 | 100.00 | 0.93 | non | oui | 2026-11-04T06:04:49.862978Z |
| docvqa | vision | 0.97 | 1.00 | 0.97 | non | oui | 2026-11-04T06:04:49.889574Z |
| dynamath | math | 0.79 | 1.00 | 0.79 | non | oui | 2026-11-04T06:04:49.973000Z |
| embspatialbench | vision | 0.83 | 100.00 | 0.83 | non | oui | 2026-11-04T06:04:49.993051Z |
| erqa | vision | 0.48 | 1.00 | 0.48 | non | oui | 2026-11-04T06:04:50.003548Z |
| hallusion-bench | vision | 0.66 | 1.00 | 0.66 | non | oui | 2026-11-04T06:04:49.869656Z |
| hr-bench-4k | vision | 0.82 | 1.00 | 0.82 | non | oui | 2026-11-04T06:04:49.879042Z |
| infovqa | vision | 0.86 | 1.00 | 0.86 | non | oui | 2026-11-04T06:04:49.923064Z |
| lvbench | vision | 0.41 | 1.00 | 0.41 | non | oui | 2026-11-04T06:04:50.053297Z |
| mathvision | math | 0.52 | 1.00 | 0.52 | non | oui | 2026-11-04T06:04:49.952432Z |
| mathvista-mini | math | 0.84 | 1.00 | 0.84 | non | oui | 2026-11-04T06:04:49.963018Z |
| mlvu-m | general | 0.71 | 1.00 | 0.71 | non | oui | 2026-11-04T06:04:50.060766Z |
| mm-if-eval | multimodal | 0.73 | 1.00 | 0.73 | non | oui | 2026-11-04T06:04:49.843061Z |
| mmbench-v1.1 | vision | 0.86 | 1.00 | 0.86 | non | oui | 2026-11-04T06:04:49.823225Z |
| mme | vision | 0.87 | 1.00 | 0.87 | non | oui | 2026-11-04T06:04:49.793770Z |
| mmmu-(val) | vision | 0.56 | 1.00 | 0.56 | non | oui | 2026-11-04T06:04:49.833579Z |
| mmmu-pro | vision | 0.46 | 1.00 | 0.46 | non | oui | 2026-11-04T06:04:49.836814Z |
| mmstar | vision | 0.74 | 1.00 | 0.74 | non | oui | 2026-11-04T06:04:49.801119Z |
| mmvu | vision | 0.53 | 1.00 | 0.53 | non | oui | 2026-11-04T06:04:50.064026Z |
| motionbench | vision | 0.64 | 1.00 | 0.64 | non | oui | 2026-11-04T06:04:50.068957Z |
| muirbench | vision | 0.66 | 1.00 | 0.66 | non | oui | 2026-11-04T06:04:49.851320Z |
| mvbench | vision | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:50.033288Z |
| ocrbench | vision | 0.86 | 1.00 | 0.86 | non | oui | 2026-11-04T06:04:49.896678Z |
| ocrbench-v2-(en) | vision | 0.64 | 1.00 | 0.64 | non | oui | 2026-11-04T06:04:49.902857Z |
| pope | vision | 0.89 | 1.00 | 0.89 | non | oui | 2026-11-04T06:04:49.873127Z |
| realworldqa | vision | 0.79 | 1.00 | 0.79 | non | oui | 2026-11-04T06:04:49.885352Z |
| refcoco-avg | vision | 0.91 | 100.00 | 0.91 | non | oui | 2026-11-04T06:04:49.980040Z |
| refspatialbench | vision | 0.50 | 100.00 | 0.50 | non | oui | 2026-11-04T06:04:49.984670Z |
| robospatialhome | vision | 0.63 | 100.00 | 0.63 | non | oui | 2026-11-04T06:04:50.013214Z |
| screenspot-pro | vision | 0.56 | 1.00 | 0.56 | non | oui | 2026-11-04T06:04:50.023532Z |
| simplevqa | vision | 0.36 | 100.00 | 0.36 | non | oui | 2026-11-04T06:04:49.810344Z |
| textvqa | vision | 0.88 | 1.00 | 0.88 | non | oui | 2026-11-04T06:04:49.892952Z |
| videomme-w-o-sub. | multimodal | 0.63 | 1.00 | 0.63 | non | oui | 2026-11-04T06:04:50.083143Z |
| videomme-w-sub. | vision | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:50.093103Z |
| videommmu | multimodal | 0.50 | 1.00 | 0.50 | non | oui | 2026-11-04T06:04:50.103136Z |
| we-math | math | 0.69 | 1.00 | 0.69 | non | oui | 2026-11-04T06:04:49.969597Z |
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