Stability AI · llama
Can I run stable-beluga?
Llama 2 based model fine tuned on an Orca-style dataset. Originally called Free Willy.
7B parameters4K contexttextSTABLE BELUGA NON-COMMERCIAL COMMUNITY LICENSE AGREEMENT18 builds
stable-beluga system requirements
Memory needed at 4K context, counting model weights, KV cache, compute buffers and runtime overhead. Download sizes come from the official Ollama registry.
| Build | Quantization | Download | Memory needed |
|---|---|---|---|
| stable-beluga:7b | Q4_0 | 3.6 GB | ~5.5 GB(est.) |
| stable-beluga:7b-q4_K_M | Q4_K_M | 3.8 GB | ~5.8 GB(est.) |
| stable-beluga:7b-q5_K_M | Q5_K_M | 4.5 GB | ~6.8 GB(est.) |
| stable-beluga:7b-q6_K | Q6_K | 5.1 GB | ~7.8 GB(est.) |
| stable-beluga:7b-q8_0 | Q8_0 | 6.7 GB | ~9.9 GB(est.) |
| stable-beluga:13b | Q4_0 | 6.9 GB | ~10.2 GB(est.) |
| stable-beluga:13b-q4_K_M | Q4_K_M | 7.3 GB | ~10.9 GB(est.) |
| stable-beluga:13b-q5_K_M | Q5_K_M | 8.6 GB | ~12.7 GB(est.) |
Figures marked (est.) approximate the KV cache because this model's architecture details are not published. They are indicative rather than exact.
How to run stable-beluga locally
- 1.Install Ollama for Windows, macOS or Linux.
- 2.Run this in a terminal:
ollama run stable-beluga:7b - 3.The weights download on first run, then an interactive prompt opens.
Frequently asked questions about stable-beluga
- How much VRAM do I need to run stable-beluga?
- At 4K context, the smallest published build of stable-beluga (Q4_0) needs roughly 5.5 GB of GPU memory once weights, KV cache, compute buffers and runtime overhead are counted. Bigger quantizations and longer context windows need more. It can also run partly or entirely on the CPU using system RAM, more slowly.
- Can I run stable-beluga without a dedicated GPU?
- Yes, but slowly. Without a GPU the model runs on the CPU using system RAM, which usually means a few tokens per second rather than dozens. You would need at least 5.5 GB of free RAM for the smallest build.
- How big is the stable-beluga download?
- The smallest published build is 3.6 GB. There are 18 builds in total, the largest being 128.5 GB. Leave some extra free disk space beyond the download itself.
- How do I run stable-beluga locally?
- Install Ollama, then run "ollama run stable-beluga:7b" in a terminal. The weights download on first use and an interactive session opens.
Will stable-beluga run on your PC?
Scan your hardware once and get a verdict for this model and every build — free, no account.
Model data from the official Ollama registry · verified 13/09/2026