All tasks / Text & knowledge / Chat & assistants
Best local model for chat & assistants on a 24 GB GPU
Open-weights models whose approximate minimum GPU memory is 24 GB or less, ranked from public benchmarks. Memory figures are approximate; quantised builds can need less.
ATop ranked: Gemma 4 (31B) (Google DeepMind), 1,452 on LMArena Text Arena (overall, style control).10 options listed · data verified 19 Oct 2026
Ranked list
| Tier | Name | Approach | Runs | Rated by | Price | Skill |
|---|---|---|---|---|---|---|
| A |  Gemma 4 (31B)Google DeepMind | Models | Local | 1,452 on LMArena Text Arena (overall, style control) | ~19 GB VRAM | — |
| A |  Qwen3.8-27BAlibaba (Qwen) | Models | Local | 1,438 on LMArena Text Arena (overall, style control) | ~17 GB VRAM | — |
| A |  Muse Glimmer 30BMeta | Models | Local | 1,425 on LMArena Text Arena (overall, style control) | ~19 GB VRAM | — |
| B |  Nemotron 3.5 Lightning 30B-A3BNVIDIA | Models | Local | 1,351 on LMArena Text Arena (overall, style control) | ~20 GB VRAM | — |
| C |  Granite 4.2 30BIBM | Models | Local | 1,340 on LMArena Text Arena (overall, style control) | ~18 GB VRAM | — |
| C |  Olmo 3.1 32BAllen Institute for AI (Ai2) | Models | Local | 1,329 on LMArena Text Arena (overall, style control) | ~20 GB VRAM | — |
| C |  gpt-oss-20bOpenAI | Models | Local | 1,317 on LMArena Text Arena (overall, style control) | ~16 GB VRAM | — |
| C |  K2-Horizon-MoVA-36B-A4BMBZUAI Institute of Foundation Models | Models | Local | 25 on Artificial Analysis Intelligence Index | ~23 GB VRAM | — |
| C | Nemotron 3.5 Bolt 74B-A9B NVIDIA | Models | Local | 21 on Artificial Analysis Intelligence Index | ~24 GB VRAM | — |
| C |  Ministral 3 14B (2512)Mistral AI | Models | Local | 6 on Artificial Analysis Intelligence Index | ~9 GB VRAM | — |
Tiers come only from public benchmarks (LMArena Text Arena (overall, style control), Artificial Analysis Intelligence Index); tools without a result of their own inherit the tier of the best model they offer. How we rank.