Model field guide
granite-3.3-8b-instruct
42 of 44 hardware configurations have a supported memory fit at one or more recorded quantizations. Runtime compatibility still needs verification.
Total parameters8.17B
Active parameters8.17B
Recorded context limit131,072
Architecturedense
Verified 2026-10-22 · license metadata: apache-2.0 · commercial use: yes (review the terms).
Actual files, not a size shortcut.
| Quantization | Total file size | Source |
|---|---|---|
| Q4_K_M | 4.603 GiB4,942,873,344 bytes; 1 file(s)Sum of exact file byte sizes in the repository listing at the recorded commit; every declared shard is present. Auxiliary vision projector and speculative-decoding files are excluded. Labels follow publisher filenames; mixed-precision expert tensors may remain in a differently labeled file, so the estimator uses measured bytes rather than assumed bits per weight. | Pinned file recordVerified 2026-10-22 |
| Q5_K_M | 5.399 GiB5,797,462,784 bytes; 1 file(s)Sum of exact file byte sizes in the repository listing at the recorded commit; every declared shard is present. Auxiliary vision projector and speculative-decoding files are excluded. Labels follow publisher filenames; mixed-precision expert tensors may remain in a differently labeled file, so the estimator uses measured bytes rather than assumed bits per weight. | Pinned file recordVerified 2026-10-22 |
| Q8_0 | 8.088 GiB8,684,264,992 bytes; 1 file(s)Sum of exact file byte sizes in the repository listing at the recorded commit; every declared shard is present. Auxiliary vision projector and speculative-decoding files are excluded. Labels follow publisher filenames; mixed-precision expert tensors may remain in a differently labeled file, so the estimator uses measured bytes rather than assumed bits per weight. | Pinned file recordVerified 2026-10-22 |
| FP16 | 15.222 GiB16,344,139,552 bytes; 1 file(s)Sum of exact file byte sizes in the repository listing at the recorded commit; every declared shard is present. Auxiliary vision projector and speculative-decoding files are excluded. Labels follow publisher filenames; mixed-precision expert tensors may remain in a differently labeled file, so the estimator uses measured bytes rather than assumed bits per weight. | Pinned file recordVerified 2026-10-22 |
The hardware picture.
Q4_K_M: hardware comparison
| Hardware | Usable memory | Required memory | Fit estimate | Estimated generation |
|---|---|---|---|---|
| Mac mini M6 (12-core GPU) 16GB | 12.0 GiB | 6.23 GiB | fits | 12.4–24.8 estimated tokens/s |
| Mac mini M6 (12-core GPU) 24GB | 18.0 GiB | 6.23 GiB | fits | 13.8–27.5 estimated tokens/s |
| Mac mini M6 (12-core GPU) 32GB | 24.0 GiB | 6.23 GiB | fits | 13.8–27.5 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 24GB | 18.0 GiB | 6.23 GiB | fits | 24.8–49.7 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 48GB | 36.0 GiB | 6.23 GiB | fits | 24.8–49.7 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 64GB | 48.0 GiB | 6.23 GiB | fits | 24.8–49.7 estimated tokens/s |
| Mac Studio M5 Max (32-core GPU) 36GB | 27.0 GiB | 6.23 GiB | fits | 37.2–74.5 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 64GB | 48.0 GiB | 6.23 GiB | fits | 49.7–99.4 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 128GB | 96.0 GiB | 6.23 GiB | fits | 49.7–99.4 estimated tokens/s |
| Mac Studio M5 Ultra (30-core CPU / 64-core GPU) 96GB | 72.0 GiB | 6.23 GiB | fits | 97.1–194.2 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 256GB | 192.0 GiB | 6.23 GiB | fits | 97.1–194.2 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 512GB | 384.0 GiB | 6.23 GiB | fits | 97.1–194.2 estimated tokens/s |
| MacBook Pro 14-inch M5 (10-core GPU) 24GB | 18.0 GiB | 6.23 GiB | fits | 12.4–24.8 estimated tokens/s |
| MacBook Pro 14-inch M5 Pro (20-core GPU) 48GB | 36.0 GiB | 6.23 GiB | fits | 24.8–49.7 estimated tokens/s |
| MacBook Pro 16-inch M5 Max (40-core GPU) 128GB | 96.0 GiB | 6.23 GiB | fits | 49.7–99.4 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 16GB | 12.0 GiB | 6.23 GiB | fits | 9.7–19.4 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 24GB | 18.0 GiB | 6.23 GiB | fits | 9.7–19.4 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 32GB | 24.0 GiB | 6.23 GiB | fits | 9.7–19.4 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 24GB | 18.0 GiB | 6.23 GiB | fits | 22.1–44.2 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 48GB | 36.0 GiB | 6.23 GiB | fits | 22.1–44.2 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 64GB | 48.0 GiB | 6.23 GiB | fits | 22.1–44.2 estimated tokens/s |
| Mac Studio (2025) M4 Max (32-core GPU) 36GB | 27.0 GiB | 6.23 GiB | fits | 33.2–66.4 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 64GB | 48.0 GiB | 6.23 GiB | fits | 44.2–88.4 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 128GB | 96.0 GiB | 6.23 GiB | fits | 44.2–88.4 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (28-core CPU / 60-core GPU) 96GB | 72.0 GiB | 6.23 GiB | fits | 66.3–132.6 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (32-core CPU / 80-core GPU) 256GB | 192.0 GiB | 6.23 GiB | fits | 66.3–132.6 estimated tokens/s |
| NVIDIA GeForce RTX 3060 12GB | 10.8 GiB | 6.23 GiB | fits | 29.1–58.3 estimated tokens/s |
| NVIDIA GeForce RTX 3090 24GB | 21.6 GiB | 6.23 GiB | fits | 75.7–151.5 estimated tokens/s |
| NVIDIA GeForce RTX 4060 Ti 16GB | 14.4 GiB | 6.23 GiB | fits | 23.3–46.6 estimated tokens/s |
| NVIDIA GeForce RTX 4070 12GB GDDR6X | 10.8 GiB | 6.23 GiB | fits | 40.8–81.6 estimated tokens/s |
| NVIDIA GeForce RTX 4080 16GB | 14.4 GiB | 6.23 GiB | fits | 58.0–116.0 estimated tokens/s |
| NVIDIA GeForce RTX 4090 24GB | 21.6 GiB | 6.23 GiB | fits | 81.6–163.1 estimated tokens/s |
| NVIDIA GeForce RTX 5060 Ti 16GB | 14.4 GiB | 6.23 GiB | fits | 36.3–72.5 estimated tokens/s |
| NVIDIA GeForce RTX 5070 12GB | 10.8 GiB | 6.23 GiB | fits | 54.4–108.8 estimated tokens/s |
| NVIDIA GeForce RTX 5070 Ti 16GB | 14.4 GiB | 6.23 GiB | fits | 72.5–145.0 estimated tokens/s |
| NVIDIA GeForce RTX 5080 16GB | 14.4 GiB | 6.23 GiB | fits | 77.7–155.4 estimated tokens/s |
| NVIDIA GeForce RTX 5090 32GB | 28.8 GiB | 6.23 GiB | fits | 145.0–290.0 estimated tokens/s |
| NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB | 86.4 GiB | 6.23 GiB | fits | 145.0–290.0 estimated tokens/s |
| NVIDIA RTX 6000 Ada Generation 48GB | 43.2 GiB | 6.23 GiB | fits | 77.7–155.4 estimated tokens/s |
| Intel Arc B580 12GB | 10.8 GiB | 6.23 GiB | fits | 36.9–73.8 estimated tokens/s |
| AMD Ryzen AI Halo, Ryzen AI Max+ 395, 128GB | 89.6 GiB | 6.23 GiB | fits | 20.7–41.4 estimated tokens/s |
| NVIDIA DGX Spark 128GB / 4TB | 96.0 GiB | 6.23 GiB | fits | 22.1–44.2 estimated tokens/s |
| 2 × GeForce RTX 3090 24GB (conditional 48GB aggregate) | 40.8 GiB | 6.23 GiB | Needs verification | Estimated speed unavailable |
| 2 × GeForce RTX 5090 32GB (conditional 64GB aggregate) | 54.4 GiB | 6.23 GiB | Needs verification | Estimated speed unavailable |
Q5_K_M: hardware comparison
| Hardware | Usable memory | Required memory | Fit estimate | Estimated generation |
|---|---|---|---|---|
| Mac mini M6 (12-core GPU) 16GB | 12.0 GiB | 7.02 GiB | fits | 10.6–21.1 estimated tokens/s |
| Mac mini M6 (12-core GPU) 24GB | 18.0 GiB | 7.02 GiB | fits | 11.7–23.5 estimated tokens/s |
| Mac mini M6 (12-core GPU) 32GB | 24.0 GiB | 7.02 GiB | fits | 11.7–23.5 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 24GB | 18.0 GiB | 7.02 GiB | fits | 21.2–42.4 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 48GB | 36.0 GiB | 7.02 GiB | fits | 21.2–42.4 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 64GB | 48.0 GiB | 7.02 GiB | fits | 21.2–42.4 estimated tokens/s |
| Mac Studio M5 Max (32-core GPU) 36GB | 27.0 GiB | 7.02 GiB | fits | 31.7–63.5 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 64GB | 48.0 GiB | 7.02 GiB | fits | 42.4–84.7 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 128GB | 96.0 GiB | 7.02 GiB | fits | 42.4–84.7 estimated tokens/s |
| Mac Studio M5 Ultra (30-core CPU / 64-core GPU) 96GB | 72.0 GiB | 7.02 GiB | fits | 82.8–165.6 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 256GB | 192.0 GiB | 7.02 GiB | fits | 82.8–165.6 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 512GB | 384.0 GiB | 7.02 GiB | fits | 82.8–165.6 estimated tokens/s |
| MacBook Pro 14-inch M5 (10-core GPU) 24GB | 18.0 GiB | 7.02 GiB | fits | 10.6–21.1 estimated tokens/s |
| MacBook Pro 14-inch M5 Pro (20-core GPU) 48GB | 36.0 GiB | 7.02 GiB | fits | 21.2–42.4 estimated tokens/s |
| MacBook Pro 16-inch M5 Max (40-core GPU) 128GB | 96.0 GiB | 7.02 GiB | fits | 42.4–84.7 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 16GB | 12.0 GiB | 7.02 GiB | fits | 8.3–16.6 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 24GB | 18.0 GiB | 7.02 GiB | fits | 8.3–16.6 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 32GB | 24.0 GiB | 7.02 GiB | fits | 8.3–16.6 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 24GB | 18.0 GiB | 7.02 GiB | fits | 18.8–37.7 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 48GB | 36.0 GiB | 7.02 GiB | fits | 18.8–37.7 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 64GB | 48.0 GiB | 7.02 GiB | fits | 18.8–37.7 estimated tokens/s |
| Mac Studio (2025) M4 Max (32-core GPU) 36GB | 27.0 GiB | 7.02 GiB | fits | 28.3–56.6 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 64GB | 48.0 GiB | 7.02 GiB | fits | 37.7–75.3 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 128GB | 96.0 GiB | 7.02 GiB | fits | 37.7–75.3 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (28-core CPU / 60-core GPU) 96GB | 72.0 GiB | 7.02 GiB | fits | 56.5–113.0 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (32-core CPU / 80-core GPU) 256GB | 192.0 GiB | 7.02 GiB | fits | 56.5–113.0 estimated tokens/s |
| NVIDIA GeForce RTX 3060 12GB | 10.8 GiB | 7.02 GiB | fits | 24.8–49.7 estimated tokens/s |
| NVIDIA GeForce RTX 3090 24GB | 21.6 GiB | 7.02 GiB | fits | 64.6–129.2 estimated tokens/s |
| NVIDIA GeForce RTX 4060 Ti 16GB | 14.4 GiB | 7.02 GiB | fits | 19.9–39.7 estimated tokens/s |
| NVIDIA GeForce RTX 4070 12GB GDDR6X | 10.8 GiB | 7.02 GiB | fits | 34.8–69.5 estimated tokens/s |
| NVIDIA GeForce RTX 4080 16GB | 14.4 GiB | 7.02 GiB | fits | 49.5–98.9 estimated tokens/s |
| NVIDIA GeForce RTX 4090 24GB | 21.6 GiB | 7.02 GiB | fits | 69.5–139.1 estimated tokens/s |
| NVIDIA GeForce RTX 5060 Ti 16GB | 14.4 GiB | 7.02 GiB | fits | 30.9–61.8 estimated tokens/s |
| NVIDIA GeForce RTX 5070 12GB | 10.8 GiB | 7.02 GiB | fits | 46.4–92.7 estimated tokens/s |
| NVIDIA GeForce RTX 5070 Ti 16GB | 14.4 GiB | 7.02 GiB | fits | 61.8–123.6 estimated tokens/s |
| NVIDIA GeForce RTX 5080 16GB | 14.4 GiB | 7.02 GiB | fits | 66.2–132.5 estimated tokens/s |
| NVIDIA GeForce RTX 5090 32GB | 28.8 GiB | 7.02 GiB | fits | 123.6–247.3 estimated tokens/s |
| NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB | 86.4 GiB | 7.02 GiB | fits | 123.6–247.3 estimated tokens/s |
| NVIDIA RTX 6000 Ada Generation 48GB | 43.2 GiB | 7.02 GiB | fits | 66.2–132.5 estimated tokens/s |
| Intel Arc B580 12GB | 10.8 GiB | 7.02 GiB | fits | 31.5–62.9 estimated tokens/s |
| AMD Ryzen AI Halo, Ryzen AI Max+ 395, 128GB | 89.6 GiB | 7.02 GiB | fits | 17.7–35.3 estimated tokens/s |
| NVIDIA DGX Spark 128GB / 4TB | 96.0 GiB | 7.02 GiB | fits | 18.8–37.7 estimated tokens/s |
| 2 × GeForce RTX 3090 24GB (conditional 48GB aggregate) | 40.8 GiB | 7.02 GiB | Needs verification | Estimated speed unavailable |
| 2 × GeForce RTX 5090 32GB (conditional 64GB aggregate) | 54.4 GiB | 7.02 GiB | Needs verification | Estimated speed unavailable |
Q8_0: hardware comparison
| Hardware | Usable memory | Required memory | Fit estimate | Estimated generation |
|---|---|---|---|---|
| Mac mini M6 (12-core GPU) 16GB | 12.0 GiB | 9.71 GiB | fits | 7.0–14.1 estimated tokens/s |
| Mac mini M6 (12-core GPU) 24GB | 18.0 GiB | 9.71 GiB | fits | 7.8–15.7 estimated tokens/s |
| Mac mini M6 (12-core GPU) 32GB | 24.0 GiB | 9.71 GiB | fits | 7.8–15.7 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 24GB | 18.0 GiB | 9.71 GiB | fits | 14.1–28.3 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 48GB | 36.0 GiB | 9.71 GiB | fits | 14.1–28.3 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 64GB | 48.0 GiB | 9.71 GiB | fits | 14.1–28.3 estimated tokens/s |
| Mac Studio M5 Max (32-core GPU) 36GB | 27.0 GiB | 9.71 GiB | fits | 21.2–42.4 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 64GB | 48.0 GiB | 9.71 GiB | fits | 28.3–56.6 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 128GB | 96.0 GiB | 9.71 GiB | fits | 28.3–56.6 estimated tokens/s |
| Mac Studio M5 Ultra (30-core CPU / 64-core GPU) 96GB | 72.0 GiB | 9.71 GiB | fits | 55.3–110.5 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 256GB | 192.0 GiB | 9.71 GiB | fits | 55.3–110.5 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 512GB | 384.0 GiB | 9.71 GiB | fits | 55.3–110.5 estimated tokens/s |
| MacBook Pro 14-inch M5 (10-core GPU) 24GB | 18.0 GiB | 9.71 GiB | fits | 7.0–14.1 estimated tokens/s |
| MacBook Pro 14-inch M5 Pro (20-core GPU) 48GB | 36.0 GiB | 9.71 GiB | fits | 14.1–28.3 estimated tokens/s |
| MacBook Pro 16-inch M5 Max (40-core GPU) 128GB | 96.0 GiB | 9.71 GiB | fits | 28.3–56.6 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 16GB | 12.0 GiB | 9.71 GiB | fits | 5.5–11.1 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 24GB | 18.0 GiB | 9.71 GiB | fits | 5.5–11.1 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 32GB | 24.0 GiB | 9.71 GiB | fits | 5.5–11.1 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 24GB | 18.0 GiB | 9.71 GiB | fits | 12.6–25.1 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 48GB | 36.0 GiB | 9.71 GiB | fits | 12.6–25.1 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 64GB | 48.0 GiB | 9.71 GiB | fits | 12.6–25.1 estimated tokens/s |
| Mac Studio (2025) M4 Max (32-core GPU) 36GB | 27.0 GiB | 9.71 GiB | fits | 18.9–37.8 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 64GB | 48.0 GiB | 9.71 GiB | fits | 25.1–50.3 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 128GB | 96.0 GiB | 9.71 GiB | fits | 25.1–50.3 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (28-core CPU / 60-core GPU) 96GB | 72.0 GiB | 9.71 GiB | fits | 37.7–75.4 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (32-core CPU / 80-core GPU) 256GB | 192.0 GiB | 9.71 GiB | fits | 37.7–75.4 estimated tokens/s |
| NVIDIA GeForce RTX 3060 12GB | 10.8 GiB | 9.71 GiB | fits | 16.6–33.2 estimated tokens/s |
| NVIDIA GeForce RTX 3090 24GB | 21.6 GiB | 9.71 GiB | fits | 43.1–86.2 estimated tokens/s |
| NVIDIA GeForce RTX 4060 Ti 16GB | 14.4 GiB | 9.71 GiB | fits | 13.3–26.5 estimated tokens/s |
| NVIDIA GeForce RTX 4070 12GB GDDR6X | 10.8 GiB | 9.71 GiB | fits | 23.2–46.4 estimated tokens/s |
| NVIDIA GeForce RTX 4080 16GB | 14.4 GiB | 9.71 GiB | fits | 33.0–66.0 estimated tokens/s |
| NVIDIA GeForce RTX 4090 24GB | 21.6 GiB | 9.71 GiB | fits | 46.4–92.9 estimated tokens/s |
| NVIDIA GeForce RTX 5060 Ti 16GB | 14.4 GiB | 9.71 GiB | fits | 20.6–41.3 estimated tokens/s |
| NVIDIA GeForce RTX 5070 12GB | 10.8 GiB | 9.71 GiB | fits | 31.0–61.9 estimated tokens/s |
| NVIDIA GeForce RTX 5070 Ti 16GB | 14.4 GiB | 9.71 GiB | fits | 41.3–82.5 estimated tokens/s |
| NVIDIA GeForce RTX 5080 16GB | 14.4 GiB | 9.71 GiB | fits | 44.2–88.4 estimated tokens/s |
| NVIDIA GeForce RTX 5090 32GB | 28.8 GiB | 9.71 GiB | fits | 82.5–165.1 estimated tokens/s |
| NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB | 86.4 GiB | 9.71 GiB | fits | 82.5–165.1 estimated tokens/s |
| NVIDIA RTX 6000 Ada Generation 48GB | 43.2 GiB | 9.71 GiB | fits | 44.2–88.4 estimated tokens/s |
| Intel Arc B580 12GB | 10.8 GiB | 9.71 GiB | fits | 21.0–42.0 estimated tokens/s |
| AMD Ryzen AI Halo, Ryzen AI Max+ 395, 128GB | 89.6 GiB | 9.71 GiB | fits | 11.8–23.6 estimated tokens/s |
| NVIDIA DGX Spark 128GB / 4TB | 96.0 GiB | 9.71 GiB | fits | 12.6–25.1 estimated tokens/s |
| 2 × GeForce RTX 3090 24GB (conditional 48GB aggregate) | 40.8 GiB | 9.71 GiB | Needs verification | Estimated speed unavailable |
| 2 × GeForce RTX 5090 32GB (conditional 64GB aggregate) | 54.4 GiB | 9.71 GiB | Needs verification | Estimated speed unavailable |
FP16: hardware comparison
| Hardware | Usable memory | Required memory | Fit estimate | Estimated generation |
|---|---|---|---|---|
| Mac mini M6 (12-core GPU) 16GB | 12.0 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| Mac mini M6 (12-core GPU) 24GB | 18.0 GiB | 16.85 GiB | borderline | 4.2–8.3 estimated tokens/s |
| Mac mini M6 (12-core GPU) 32GB | 24.0 GiB | 16.85 GiB | fits | 4.2–8.3 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 24GB | 18.0 GiB | 16.85 GiB | borderline | 7.5–15.0 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 48GB | 36.0 GiB | 16.85 GiB | fits | 7.5–15.0 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 64GB | 48.0 GiB | 16.85 GiB | fits | 7.5–15.0 estimated tokens/s |
| Mac Studio M5 Max (32-core GPU) 36GB | 27.0 GiB | 16.85 GiB | fits | 11.3–22.5 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 64GB | 48.0 GiB | 16.85 GiB | fits | 15.0–30.1 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 128GB | 96.0 GiB | 16.85 GiB | fits | 15.0–30.1 estimated tokens/s |
| Mac Studio M5 Ultra (30-core CPU / 64-core GPU) 96GB | 72.0 GiB | 16.85 GiB | fits | 29.4–58.7 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 256GB | 192.0 GiB | 16.85 GiB | fits | 29.4–58.7 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 512GB | 384.0 GiB | 16.85 GiB | fits | 29.4–58.7 estimated tokens/s |
| MacBook Pro 14-inch M5 (10-core GPU) 24GB | 18.0 GiB | 16.85 GiB | borderline | 3.7–7.5 estimated tokens/s |
| MacBook Pro 14-inch M5 Pro (20-core GPU) 48GB | 36.0 GiB | 16.85 GiB | fits | 7.5–15.0 estimated tokens/s |
| MacBook Pro 16-inch M5 Max (40-core GPU) 128GB | 96.0 GiB | 16.85 GiB | fits | 15.0–30.1 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 16GB | 12.0 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| Mac mini (2024) M4 (10-core GPU) 24GB | 18.0 GiB | 16.85 GiB | borderline | 2.9–5.9 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 32GB | 24.0 GiB | 16.85 GiB | fits | 2.9–5.9 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 24GB | 18.0 GiB | 16.85 GiB | borderline | 6.7–13.4 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 48GB | 36.0 GiB | 16.85 GiB | fits | 6.7–13.4 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 64GB | 48.0 GiB | 16.85 GiB | fits | 6.7–13.4 estimated tokens/s |
| Mac Studio (2025) M4 Max (32-core GPU) 36GB | 27.0 GiB | 16.85 GiB | fits | 10.0–20.1 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 64GB | 48.0 GiB | 16.85 GiB | fits | 13.4–26.7 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 128GB | 96.0 GiB | 16.85 GiB | fits | 13.4–26.7 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (28-core CPU / 60-core GPU) 96GB | 72.0 GiB | 16.85 GiB | fits | 20.0–40.1 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (32-core CPU / 80-core GPU) 256GB | 192.0 GiB | 16.85 GiB | fits | 20.0–40.1 estimated tokens/s |
| NVIDIA GeForce RTX 3060 12GB | 10.8 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 3090 24GB | 21.6 GiB | 16.85 GiB | fits | 22.9–45.8 estimated tokens/s |
| NVIDIA GeForce RTX 4060 Ti 16GB | 14.4 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 4070 12GB GDDR6X | 10.8 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 4080 16GB | 14.4 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 4090 24GB | 21.6 GiB | 16.85 GiB | fits | 24.7–49.3 estimated tokens/s |
| NVIDIA GeForce RTX 5060 Ti 16GB | 14.4 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 5070 12GB | 10.8 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 5070 Ti 16GB | 14.4 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 5080 16GB | 14.4 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| NVIDIA GeForce RTX 5090 32GB | 28.8 GiB | 16.85 GiB | fits | 43.9–87.7 estimated tokens/s |
| NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB | 86.4 GiB | 16.85 GiB | fits | 43.9–87.7 estimated tokens/s |
| NVIDIA RTX 6000 Ada Generation 48GB | 43.2 GiB | 16.85 GiB | fits | 23.5–47.0 estimated tokens/s |
| Intel Arc B580 12GB | 10.8 GiB | 16.85 GiB | doesn't fit | Estimated speed unavailable |
| AMD Ryzen AI Halo, Ryzen AI Max+ 395, 128GB | 89.6 GiB | 16.85 GiB | fits | 6.3–12.5 estimated tokens/s |
| NVIDIA DGX Spark 128GB / 4TB | 96.0 GiB | 16.85 GiB | fits | 6.7–13.4 estimated tokens/s |
| 2 × GeForce RTX 3090 24GB (conditional 48GB aggregate) | 40.8 GiB | 16.85 GiB | Needs verification | Estimated speed unavailable |
| 2 × GeForce RTX 5090 32GB (conditional 64GB aggregate) | 54.4 GiB | 16.85 GiB | Needs verification | Estimated speed unavailable |
A “fits” result checks estimated memory only. “Needs verification” preserves missing cache, vision, or placement information. All speed figures are estimated from the published bandwidth formula.
Field-level sources and verification dates
- id: Official source · verified 2026-10-22. Lowercase URL slug derived from the upstream repository ID; baseModel retains the exact Hugging Face identifier.
- name: Official source · verified 2026-10-22
- baseModel: Official source · verified 2026-10-22
- downloads: Official source · verified 2026-10-22
- tasks: Official source · verified 2026-10-22. Chat denotes a text-generating model, including base models that may require a chat template or instruction tuning. Code label is inferred from the repository model-family name. Vision uses pipeline metadata or vision_config.
- license: Official source · verified 2026-10-22
- commercialUse: Official source · verified 2026-10-22. Official upstream organization metadata identifies a permissive standard license. Commercial use is permitted subject to the license terms, notices, and applicable restrictions; this is a metadata-based classification, not an individualized legal approval.
- license_url: Official source · verified 2026-10-22. No recognized standalone license file; linked pinned model card.
- architecture: Official source · verified 2026-10-22. Classification from model_type and expert/state-space configuration fields (granite); hybrid classification takes precedence over expert count.
- parametersB: Official source · verified 2026-10-22. Exact safetensors parameter total reported by the Hugging Face API, divided by 1e9.
- activeParametersB: Official source · verified 2026-10-22. Dense model: all parameters are included. This is not a runtime FLOP measurement.
- contextLength: Official source · verified 2026-10-22. Configured maximum token positions; not a guarantee of runtime support or quality at that length. No speculative RoPE scaling applied.
- kvCache: Official source · verified 2026-10-22. Standard attention KV shape from num_hidden_layers, num_key_value_heads, and head_dim or hidden_size / num_attention_heads; FP16 cache bytes are a separate estimator choice.
- licenseStandard: Official source · verified 2026-10-22. Official standard license text supporting the permissive-license classification; model applicability is identified separately by upstream metadata.