Trouvez le meilleur LLM pour votre matériel
Sélectionnez votre GPU et RAM système pour voir les modèles classés par adéquation VRAM. Obtenez des configurations optimisées pour votre infrastructure.
Configuration matérielle
Facultatif — pour des recommandations de déploiement précises
Filtres de modèles
1B
1026B
4K
10M
| Modèle | Organisation | Paramètres | Contexte | Capacités | Quants | Adéquation |
|---|---|---|---|---|---|---|
| Kimi K2 Instruct | Moonshot AI | 1026.47B | 128K |
Code
Tool Calls
|
18 | — |
| Kimi K2.5 | Moonshot AI | 1016.23B | 256K |
Code
Thinking
Tool Calls
Vision
|
18 | — |
| DeepSeek V3.2 | DeepSeek | 685.4B | 160K |
Code
Multilingual
Thinking
Tool Calls
|
18 | — |
| DeepSeek V3.1 | DeepSeek | 684.53B | 160K |
Code
Multilingual
Thinking
Tool Calls
|
18 | — |
| Mistral Large 3 675B Instruct 2512 | Mistral AI | 675B | 288K |
Code
Multilingual
Tool Calls
|
18 | — |
| Qwen3.5 397B A17B | Qwen | 403.4B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
22 | — |
| Llama 4 Maverick 17B 128E Instruct | Meta | 396.58B | 1M |
Code
Multilingual
Tool Calls
Vision
|
18 | — |
| GLM 4.7 | Zai Org | 358.34B | 198K |
Code
Thinking
Tool Calls
|
18 | — |
| Qwen3 235B A22B | Qwen | 235.09B | 40K |
Code
Multilingual
Thinking
Tool Calls
|
17 | — |
| MiniMax M2 | MiniMax | 228.7B | 192K |
Code
Thinking
Tool Calls
|
18 | — |
| Kimi K2 Thinking | Moonshot AI | 170.27B | 256K |
Code
Thinking
Tool Calls
|
18 | — |
| Devstral 2 123B Instruct 2512 | Mistral AI | 125.03B | 256K |
Code
Multilingual
Tool Calls
|
18 | — |
| NVIDIA Nemotron 3 Super 120B A12B | NVIDIA | 123.61B | 256K |
Code
Multilingual
Thinking
Tool Calls
|
15 | — |
| Mistral Large Instruct 2411 | Mistral AI | 122.61B | 128K |
Multilingual
Tool Calls
|
14 | — |
| GPT OSS 120B | OpenAI | 120.41B | 128K |
Multilingual
Thinking
Tool Calls
|
16 | — |
| Mistral Small 4 119B 2603 | Mistral AI | 119.4B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
3 | — |
| Qwen3 Next 80B A3B Thinking | Qwen | 81.32B | 256K |
Code
Multilingual
Thinking
Tool Calls
|
18 | — |
| Qwen3 Next 80B A3B Instruct | Qwen | 81.32B | 256K |
Code
Multilingual
Tool Calls
|
18 | — |
| Qwen3 Coder Next | Qwen | 79.67B | 256K |
Code
Multilingual
Tool Calls
|
19 | — |
| Qwen2.5 72B Instruct | Qwen | 72.71B | 32K |
Code
Multilingual
Tool Calls
|
9 | — |
| DeepSeek R1 Distill Llama 70B | DeepSeek | 70.55B | 128K |
Code
Multilingual
Thinking
Tool Calls
|
19 | — |
| Llama 3.3 70B Instruct | Meta | 70B | 128K |
Code
Multilingual
Tool Calls
|
20 | — |
| Meta Llama 3.1 70B Instruct | Meta | 70B | 128K |
Code
Multilingual
Tool Calls
|
15 | — |
| Qwen3.5 35B A3B | Qwen | 35.95B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
11 | — |
| DeepSeek R1 Distill Qwen 32B | DeepSeek | 32.76B | 128K |
Code
Multilingual
Thinking
Tool Calls
|
8 | — |
| Qwen3 32B | Qwen | 32B | 40K |
Code
Multilingual
Thinking
Tool Calls
|
18 | — |
| NVIDIA Nemotron 3 Nano 30B A3B | NVIDIA | 31.58B | 256K |
Code
Multilingual
Thinking
Tool Calls
|
17 | — |
| GLM 4.7 Flash | Zai Org | 31.22B | 198K |
Code
Thinking
Tool Calls
|
17 | — |
| Qwen3.5 27B | Qwen | 27.78B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
16 | — |
| Devstral Small 2 24B Instruct 2512 | Mistral AI | 24.01B | 384K |
Code
Multilingual
Tool Calls
|
18 | — |
| Mistral Small 3.1 24B Instruct 2503 | Mistral AI | 24B | 128K |
Code
Multilingual
Tool Calls
Vision
|
18 | — |
| Mistral Small 24B Instruct 2501 | Mistral AI | 23.57B | 32K |
Code
Multilingual
Tool Calls
|
19 | — |
| GPT OSS 20B | OpenAI | 21.51B | 128K |
Multilingual
Thinking
Tool Calls
|
16 | — |
| Llama 4 Scout 17B 16E Instruct | Meta | 17B | 10M |
Code
Multilingual
Tool Calls
Vision
|
18 | — |
| DeepSeek Coder V2 Lite Instruct | DeepSeek | 15.71B | 160K |
Code
Multilingual
|
5 | — |
| DeepSeek R1 Distill Qwen 14B | DeepSeek | 14.77B | 128K |
Code
Multilingual
Thinking
Tool Calls
|
8 | — |
| Qwen2.5 14B Instruct | Qwen | 14.77B | 32K |
Code
Multilingual
Tool Calls
|
9 | — |
| Phi 4 | Microsoft | 14.66B | 16K |
Code
|
14 | — |
| Qwen3.5 9B | Qwen | 9.65B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
16 | — |
| DeepSeek R1 0528 Qwen3 8B | DeepSeek | 8.19B | 128K |
Code
Multilingual
Thinking
Tool Calls
|
18 | — |
| Meta Llama 3.1 8B Instruct | Meta | 8B | 128K |
Code
Multilingual
Tool Calls
|
19 | — |
| Qwen3 8B | Qwen | 8B | 40K |
Code
Multilingual
Thinking
Tool Calls
|
17 | — |
| DeepSeek R1 Distill Qwen 7B | DeepSeek | 7.62B | 128K |
Code
Multilingual
Thinking
Tool Calls
|
8 | — |
| Qwen2.5 7B Instruct | Qwen | 7.62B | 32K |
Code
Multilingual
Tool Calls
|
9 | — |
| Granite 4.0 Tiny Base Preview | IBM | 6.67B | 128K |
Code
Multilingual
|
15 | — |
| Qwen3.5 4B | Qwen | 4.66B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
16 | — |
| NVIDIA Nemotron 3 Nano 4B | NVIDIA | 3.97B | 256K |
Code
Thinking
Tool Calls
|
22 | — |
| Phi 3 mini 4k instruct | Microsoft | 3.82B | 4K |
Code
|
2 | — |
| Qwen3.5 2B | Qwen | 2.27B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
16 | — |
| LFM2.5 1.2B Thinking | Liquid AI | 1.17B | 125K |
Multilingual
Thinking
Tool Calls
|
6 | — |
| Qwen3.5 0.8B | Qwen | 0.87B | 256K |
Code
Multilingual
Thinking
Tool Calls
Vision
|
16 | — |
Dernière mise à jour : 20 mars 2026