~ > spark model list -- 11 models, one list
| model | file | RAM | license | tested | grounded |
|---|---|---|---|---|---|
| gemma4-26b-a4b | 15.9 GB | 19 GB | Apache 2.0 | yes | 27/30 2026-10-27 |
| gemma4-e4b | 5.0 GB | 8 GB | Apache 2.0 | yes | 27/30 2026-10-27 |
| qwen3-4b | 2.3 GB | 5 GB | Apache 2.0 | yes | 17/30 2026-10-28 |
| qwen3-5-2b | 1.3 GB | 3 GB | Apache 2.0 | yes | 15/30 2026-10-28 |
| qwen3-5-4b | 2.8 GB | 5 GB | Apache 2.0 | yes | 20/30 2026-10-28 |
| qwen3-8b | 4.7 GB | 7 GB | Apache 2.0 | yes | 22/30 2026-10-28 |
| granite-4-2-8b | 5.0 GB | 7 GB | Apache 2.0 | yes | 26/30 2026-10-28 |
| qwen3-14b | 8.4 GB | 11 GB | Apache 2.0 | yes | 19/30 2026-10-28 |
| qwen3-30b-a3b | 17.4 GB | 21 GB | Apache 2.0 | yes | 24/30 2026-10-27 |
| gemma4-e2b | 3.2 GB | 6 GB | Apache 2.0 | 21/30 2026-09-30 | |
| qwen3-coder-30b-a3b | 17.3 GB | 21 GB | Apache 2.0 |
spark model NAME serves one model. spark ember NAME adds the chat model, a bigger second one. auto picks a row tested at the prompt line under Apache-2.0 or MIT. Any other license is shown and asked about before the download. Every file is checked against its sha256. Your own rows: spark model add URL --license. A row is a pull request: docs/CONTRIBUTING.md.
source: models.env on GitHub -- this page is that file at spark v1.113.1