Autor: Jakub Rusinowski · Ostatnia aktualizacja: 6 września 2026
MIT-licensed reasoning fine-tunes that argue the case for small models: VibeThinker-3B scores 94.3 on AIME26, matching what DeepSeek V3.2 reaches at 671B. Trained with the Spectrum-to-Signal pipeline — explore solution diversity during SFT, then reinforce the correct signal with RL. Vendor-reported and contested; treat the headline comparison as a claim about a narrow benchmark family, not general capability.
| Licence | What it permits | Applies to |
|---|---|---|
MIT | Commercial use permitted Commercial use permitted. No usage restrictions beyond attribution. | VibeThinker 3B, VibeThinker 1.5B |
| VibeThinker 3B | Min 3 GB VRAM · Q4_K_M · 131,072 ctx · |
| VibeThinker 1.5B | Min 2 GB VRAM · Q4_K_M · 131,072 ctx · |
The cheapest GPU that runs VibeThinker locally (min 2 GB VRAM) is the Intel Arc B570 (10 GB).
Install Ollama then run: ollama run
Minimum VRAM: 2 GB. For best results use Q4_K_M quantization.
VibeThinker needs about 2 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: VibeThinker 3B (3 GB, Q4_K_M); VibeThinker 1.5B (2 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Yes — VibeThinker runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
Q4_K_M is the best balance of quality and VRAM for VibeThinker in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
Install Ollama, then run: ollama run . This downloads VibeThinker and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.