Model field guide
granite-4.1-3b
42 of 44 hardware configurations have a supported memory fit at one or more recorded quantizations. Runtime compatibility still needs verification.
Total parameters3.40B
Active parameters3.40B
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 | 1.955 GiB2,099,501,664 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 | 2.270 GiB2,437,012,064 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 | 3.371 GiB3,619,691,104 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 | 3.27 GiB | fits | 29.1–58.3 estimated tokens/s |
| Mac mini M6 (12-core GPU) 24GB | 18.0 GiB | 3.27 GiB | fits | 32.4–64.8 estimated tokens/s |
| Mac mini M6 (12-core GPU) 32GB | 24.0 GiB | 3.27 GiB | fits | 32.4–64.8 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 24GB | 18.0 GiB | 3.27 GiB | fits | 58.5–117.0 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 48GB | 36.0 GiB | 3.27 GiB | fits | 58.5–117.0 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 64GB | 48.0 GiB | 3.27 GiB | fits | 58.5–117.0 estimated tokens/s |
| Mac Studio M5 Max (32-core GPU) 36GB | 27.0 GiB | 3.27 GiB | fits | 87.6–175.3 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 64GB | 48.0 GiB | 3.27 GiB | fits | 117.0–234.0 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 128GB | 96.0 GiB | 3.27 GiB | fits | 117.0–234.0 estimated tokens/s |
| Mac Studio M5 Ultra (30-core CPU / 64-core GPU) 96GB | 72.0 GiB | 3.27 GiB | fits | 228.6–457.3 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 256GB | 192.0 GiB | 3.27 GiB | fits | 228.6–457.3 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 512GB | 384.0 GiB | 3.27 GiB | fits | 228.6–457.3 estimated tokens/s |
| MacBook Pro 14-inch M5 (10-core GPU) 24GB | 18.0 GiB | 3.27 GiB | fits | 29.1–58.3 estimated tokens/s |
| MacBook Pro 14-inch M5 Pro (20-core GPU) 48GB | 36.0 GiB | 3.27 GiB | fits | 58.5–117.0 estimated tokens/s |
| MacBook Pro 16-inch M5 Max (40-core GPU) 128GB | 96.0 GiB | 3.27 GiB | fits | 117.0–234.0 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 16GB | 12.0 GiB | 3.27 GiB | fits | 22.9–45.7 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 24GB | 18.0 GiB | 3.27 GiB | fits | 22.9–45.7 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 32GB | 24.0 GiB | 3.27 GiB | fits | 22.9–45.7 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 24GB | 18.0 GiB | 3.27 GiB | fits | 52.0–104.0 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 48GB | 36.0 GiB | 3.27 GiB | fits | 52.0–104.0 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 64GB | 48.0 GiB | 3.27 GiB | fits | 52.0–104.0 estimated tokens/s |
| Mac Studio (2025) M4 Max (32-core GPU) 36GB | 27.0 GiB | 3.27 GiB | fits | 78.1–156.2 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 64GB | 48.0 GiB | 3.27 GiB | fits | 104.0–208.0 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 128GB | 96.0 GiB | 3.27 GiB | fits | 104.0–208.0 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (28-core CPU / 60-core GPU) 96GB | 72.0 GiB | 3.27 GiB | fits | 156.0–312.1 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (32-core CPU / 80-core GPU) 256GB | 192.0 GiB | 3.27 GiB | fits | 156.0–312.1 estimated tokens/s |
| NVIDIA GeForce RTX 3060 12GB | 10.8 GiB | 3.27 GiB | fits | 68.6–137.2 estimated tokens/s |
| NVIDIA GeForce RTX 3090 24GB | 21.6 GiB | 3.27 GiB | fits | 178.3–356.7 estimated tokens/s |
| NVIDIA GeForce RTX 4060 Ti 16GB | 14.4 GiB | 3.27 GiB | fits | 54.9–109.7 estimated tokens/s |
| NVIDIA GeForce RTX 4070 12GB GDDR6X | 10.8 GiB | 3.27 GiB | fits | 96.0–192.0 estimated tokens/s |
| NVIDIA GeForce RTX 4080 16GB | 14.4 GiB | 3.27 GiB | fits | 136.6–273.1 estimated tokens/s |
| NVIDIA GeForce RTX 4090 24GB | 21.6 GiB | 3.27 GiB | fits | 192.0–384.1 estimated tokens/s |
| NVIDIA GeForce RTX 5060 Ti 16GB | 14.4 GiB | 3.27 GiB | fits | 85.4–170.7 estimated tokens/s |
| NVIDIA GeForce RTX 5070 12GB | 10.8 GiB | 3.27 GiB | fits | 128.0–256.1 estimated tokens/s |
| NVIDIA GeForce RTX 5070 Ti 16GB | 14.4 GiB | 3.27 GiB | fits | 170.7–341.4 estimated tokens/s |
| NVIDIA GeForce RTX 5080 16GB | 14.4 GiB | 3.27 GiB | fits | 182.9–365.8 estimated tokens/s |
| NVIDIA GeForce RTX 5090 32GB | 28.8 GiB | 3.27 GiB | fits | 341.4–682.8 estimated tokens/s |
| NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB | 86.4 GiB | 3.27 GiB | fits | 341.4–682.8 estimated tokens/s |
| NVIDIA RTX 6000 Ada Generation 48GB | 43.2 GiB | 3.27 GiB | fits | 182.9–365.8 estimated tokens/s |
| Intel Arc B580 12GB | 10.8 GiB | 3.27 GiB | fits | 86.9–173.8 estimated tokens/s |
| AMD Ryzen AI Halo, Ryzen AI Max+ 395, 128GB | 89.6 GiB | 3.27 GiB | fits | 48.8–97.5 estimated tokens/s |
| NVIDIA DGX Spark 128GB / 4TB | 96.0 GiB | 3.27 GiB | fits | 52.0–104.0 estimated tokens/s |
| 2 × GeForce RTX 3090 24GB (conditional 48GB aggregate) | 40.8 GiB | 3.27 GiB | Needs verification | Estimated speed unavailable |
| 2 × GeForce RTX 5090 32GB (conditional 64GB aggregate) | 54.4 GiB | 3.27 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 | 3.58 GiB | fits | 25.1–50.2 estimated tokens/s |
| Mac mini M6 (12-core GPU) 24GB | 18.0 GiB | 3.58 GiB | fits | 27.9–55.8 estimated tokens/s |
| Mac mini M6 (12-core GPU) 32GB | 24.0 GiB | 3.58 GiB | fits | 27.9–55.8 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 24GB | 18.0 GiB | 3.58 GiB | fits | 50.4–100.8 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 48GB | 36.0 GiB | 3.58 GiB | fits | 50.4–100.8 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 64GB | 48.0 GiB | 3.58 GiB | fits | 50.4–100.8 estimated tokens/s |
| Mac Studio M5 Max (32-core GPU) 36GB | 27.0 GiB | 3.58 GiB | fits | 75.5–151.0 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 64GB | 48.0 GiB | 3.58 GiB | fits | 100.8–201.6 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 128GB | 96.0 GiB | 3.58 GiB | fits | 100.8–201.6 estimated tokens/s |
| Mac Studio M5 Ultra (30-core CPU / 64-core GPU) 96GB | 72.0 GiB | 3.58 GiB | fits | 197.0–393.9 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 256GB | 192.0 GiB | 3.58 GiB | fits | 197.0–393.9 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 512GB | 384.0 GiB | 3.58 GiB | fits | 197.0–393.9 estimated tokens/s |
| MacBook Pro 14-inch M5 (10-core GPU) 24GB | 18.0 GiB | 3.58 GiB | fits | 25.1–50.2 estimated tokens/s |
| MacBook Pro 14-inch M5 Pro (20-core GPU) 48GB | 36.0 GiB | 3.58 GiB | fits | 50.4–100.8 estimated tokens/s |
| MacBook Pro 16-inch M5 Max (40-core GPU) 128GB | 96.0 GiB | 3.58 GiB | fits | 100.8–201.6 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 16GB | 12.0 GiB | 3.58 GiB | fits | 19.7–39.4 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 24GB | 18.0 GiB | 3.58 GiB | fits | 19.7–39.4 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 32GB | 24.0 GiB | 3.58 GiB | fits | 19.7–39.4 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 24GB | 18.0 GiB | 3.58 GiB | fits | 44.8–89.6 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 48GB | 36.0 GiB | 3.58 GiB | fits | 44.8–89.6 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 64GB | 48.0 GiB | 3.58 GiB | fits | 44.8–89.6 estimated tokens/s |
| Mac Studio (2025) M4 Max (32-core GPU) 36GB | 27.0 GiB | 3.58 GiB | fits | 67.3–134.6 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 64GB | 48.0 GiB | 3.58 GiB | fits | 89.6–179.2 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 128GB | 96.0 GiB | 3.58 GiB | fits | 89.6–179.2 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (28-core CPU / 60-core GPU) 96GB | 72.0 GiB | 3.58 GiB | fits | 134.4–268.9 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (32-core CPU / 80-core GPU) 256GB | 192.0 GiB | 3.58 GiB | fits | 134.4–268.9 estimated tokens/s |
| NVIDIA GeForce RTX 3060 12GB | 10.8 GiB | 3.58 GiB | fits | 59.1–118.2 estimated tokens/s |
| NVIDIA GeForce RTX 3090 24GB | 21.6 GiB | 3.58 GiB | fits | 153.6–307.3 estimated tokens/s |
| NVIDIA GeForce RTX 4060 Ti 16GB | 14.4 GiB | 3.58 GiB | fits | 47.3–94.5 estimated tokens/s |
| NVIDIA GeForce RTX 4070 12GB GDDR6X | 10.8 GiB | 3.58 GiB | fits | 82.7–165.4 estimated tokens/s |
| NVIDIA GeForce RTX 4080 16GB | 14.4 GiB | 3.58 GiB | fits | 117.7–235.3 estimated tokens/s |
| NVIDIA GeForce RTX 4090 24GB | 21.6 GiB | 3.58 GiB | fits | 165.4–330.9 estimated tokens/s |
| NVIDIA GeForce RTX 5060 Ti 16GB | 14.4 GiB | 3.58 GiB | fits | 73.5–147.1 estimated tokens/s |
| NVIDIA GeForce RTX 5070 12GB | 10.8 GiB | 3.58 GiB | fits | 110.3–220.6 estimated tokens/s |
| NVIDIA GeForce RTX 5070 Ti 16GB | 14.4 GiB | 3.58 GiB | fits | 147.1–294.1 estimated tokens/s |
| NVIDIA GeForce RTX 5080 16GB | 14.4 GiB | 3.58 GiB | fits | 157.6–315.1 estimated tokens/s |
| NVIDIA GeForce RTX 5090 32GB | 28.8 GiB | 3.58 GiB | fits | 294.1–588.3 estimated tokens/s |
| NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB | 86.4 GiB | 3.58 GiB | fits | 294.1–588.3 estimated tokens/s |
| NVIDIA RTX 6000 Ada Generation 48GB | 43.2 GiB | 3.58 GiB | fits | 157.6–315.1 estimated tokens/s |
| Intel Arc B580 12GB | 10.8 GiB | 3.58 GiB | fits | 74.8–149.7 estimated tokens/s |
| AMD Ryzen AI Halo, Ryzen AI Max+ 395, 128GB | 89.6 GiB | 3.58 GiB | fits | 42.0–84.0 estimated tokens/s |
| NVIDIA DGX Spark 128GB / 4TB | 96.0 GiB | 3.58 GiB | fits | 44.8–89.6 estimated tokens/s |
| 2 × GeForce RTX 3090 24GB (conditional 48GB aggregate) | 40.8 GiB | 3.58 GiB | Needs verification | Estimated speed unavailable |
| 2 × GeForce RTX 5090 32GB (conditional 64GB aggregate) | 54.4 GiB | 3.58 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 | 4.68 GiB | fits | 16.9–33.8 estimated tokens/s |
| Mac mini M6 (12-core GPU) 24GB | 18.0 GiB | 4.68 GiB | fits | 18.8–37.6 estimated tokens/s |
| Mac mini M6 (12-core GPU) 32GB | 24.0 GiB | 4.68 GiB | fits | 18.8–37.6 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 24GB | 18.0 GiB | 4.68 GiB | fits | 33.9–67.9 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 48GB | 36.0 GiB | 4.68 GiB | fits | 33.9–67.9 estimated tokens/s |
| Mac mini M5 Pro (15-core CPU / 16-core GPU) 64GB | 48.0 GiB | 4.68 GiB | fits | 33.9–67.9 estimated tokens/s |
| Mac Studio M5 Max (32-core GPU) 36GB | 27.0 GiB | 4.68 GiB | fits | 50.8–101.7 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 64GB | 48.0 GiB | 4.68 GiB | fits | 67.9–135.7 estimated tokens/s |
| Mac Studio M5 Max (40-core GPU) 128GB | 96.0 GiB | 4.68 GiB | fits | 67.9–135.7 estimated tokens/s |
| Mac Studio M5 Ultra (30-core CPU / 64-core GPU) 96GB | 72.0 GiB | 4.68 GiB | fits | 132.6–265.2 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 256GB | 192.0 GiB | 4.68 GiB | fits | 132.6–265.2 estimated tokens/s |
| Mac Studio M5 Ultra (36-core CPU / 80-core GPU) 512GB | 384.0 GiB | 4.68 GiB | fits | 132.6–265.2 estimated tokens/s |
| MacBook Pro 14-inch M5 (10-core GPU) 24GB | 18.0 GiB | 4.68 GiB | fits | 16.9–33.8 estimated tokens/s |
| MacBook Pro 14-inch M5 Pro (20-core GPU) 48GB | 36.0 GiB | 4.68 GiB | fits | 33.9–67.9 estimated tokens/s |
| MacBook Pro 16-inch M5 Max (40-core GPU) 128GB | 96.0 GiB | 4.68 GiB | fits | 67.9–135.7 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 16GB | 12.0 GiB | 4.68 GiB | fits | 13.3–26.5 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 24GB | 18.0 GiB | 4.68 GiB | fits | 13.3–26.5 estimated tokens/s |
| Mac mini (2024) M4 (10-core GPU) 32GB | 24.0 GiB | 4.68 GiB | fits | 13.3–26.5 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 24GB | 18.0 GiB | 4.68 GiB | fits | 30.2–60.3 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 48GB | 36.0 GiB | 4.68 GiB | fits | 30.2–60.3 estimated tokens/s |
| Mac mini (2024) M4 Pro (16-core GPU) 64GB | 48.0 GiB | 4.68 GiB | fits | 30.2–60.3 estimated tokens/s |
| Mac Studio (2025) M4 Max (32-core GPU) 36GB | 27.0 GiB | 4.68 GiB | fits | 45.3–90.6 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 64GB | 48.0 GiB | 4.68 GiB | fits | 60.3–120.7 estimated tokens/s |
| Mac Studio (2025) M4 Max (40-core GPU) 128GB | 96.0 GiB | 4.68 GiB | fits | 60.3–120.7 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (28-core CPU / 60-core GPU) 96GB | 72.0 GiB | 4.68 GiB | fits | 90.5–181.0 estimated tokens/s |
| Mac Studio (2025) M3 Ultra (32-core CPU / 80-core GPU) 256GB | 192.0 GiB | 4.68 GiB | fits | 90.5–181.0 estimated tokens/s |
| NVIDIA GeForce RTX 3060 12GB | 10.8 GiB | 4.68 GiB | fits | 39.8–79.6 estimated tokens/s |
| NVIDIA GeForce RTX 3090 24GB | 21.6 GiB | 4.68 GiB | fits | 103.4–206.9 estimated tokens/s |
| NVIDIA GeForce RTX 4060 Ti 16GB | 14.4 GiB | 4.68 GiB | fits | 31.8–63.7 estimated tokens/s |
| NVIDIA GeForce RTX 4070 12GB GDDR6X | 10.8 GiB | 4.68 GiB | fits | 55.7–111.4 estimated tokens/s |
| NVIDIA GeForce RTX 4080 16GB | 14.4 GiB | 4.68 GiB | fits | 79.2–158.4 estimated tokens/s |
| NVIDIA GeForce RTX 4090 24GB | 21.6 GiB | 4.68 GiB | fits | 111.4–222.8 estimated tokens/s |
| NVIDIA GeForce RTX 5060 Ti 16GB | 14.4 GiB | 4.68 GiB | fits | 49.5–99.0 estimated tokens/s |
| NVIDIA GeForce RTX 5070 12GB | 10.8 GiB | 4.68 GiB | fits | 74.3–148.5 estimated tokens/s |
| NVIDIA GeForce RTX 5070 Ti 16GB | 14.4 GiB | 4.68 GiB | fits | 99.0–198.0 estimated tokens/s |
| NVIDIA GeForce RTX 5080 16GB | 14.4 GiB | 4.68 GiB | fits | 106.1–212.2 estimated tokens/s |
| NVIDIA GeForce RTX 5090 32GB | 28.8 GiB | 4.68 GiB | fits | 198.0–396.1 estimated tokens/s |
| NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB | 86.4 GiB | 4.68 GiB | fits | 198.0–396.1 estimated tokens/s |
| NVIDIA RTX 6000 Ada Generation 48GB | 43.2 GiB | 4.68 GiB | fits | 106.1–212.2 estimated tokens/s |
| Intel Arc B580 12GB | 10.8 GiB | 4.68 GiB | fits | 50.4–100.8 estimated tokens/s |
| AMD Ryzen AI Halo, Ryzen AI Max+ 395, 128GB | 89.6 GiB | 4.68 GiB | fits | 28.3–56.6 estimated tokens/s |
| NVIDIA DGX Spark 128GB / 4TB | 96.0 GiB | 4.68 GiB | fits | 30.2–60.3 estimated tokens/s |
| 2 × GeForce RTX 3090 24GB (conditional 48GB aggregate) | 40.8 GiB | 4.68 GiB | Needs verification | Estimated speed unavailable |
| 2 × GeForce RTX 5090 32GB (conditional 64GB aggregate) | 54.4 GiB | 4.68 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.