Quick start
ollama run clef:27b-nvfp4Runnable variants
Memory is a rough file-size estimate; context length, runtime and GPU offload change it.
9 of 9 matching variants shown.9 of 9 variants shown.
| Tag | Type | Inputs | Digest | Download | Parameters | Quantization | Context | Est. memory | Run |
|---|---|---|---|---|---|---|---|---|---|
| 27b | Typelocal | Inputstext, image | Digestfdd5939e1bd9 | DownloadUnknown | Parameters27b | QuantizationUnknown | Context256K | Est. memoryUnknown | ollama run clef:27b |
| 27b-cloud | Typecloud | Inputstext, image | Digest5896d5380b6c | DownloadUnknown | Parameters27b | QuantizationUnknown | Context64K | Est. memoryUnknown | ollama run clef:27b-cloud |
| 27b-mlx-bf16 | Typelocal | Inputstext, image | Digest3f0f84ddfcf6 | Download55 GB | Parameters27b | QuantizationBF16 | Context256K | Est. memory≈ 62.5 GB | ollama run clef:27b-mlx-bf16 |
| 27b-mxfp8 | Typelocal | Inputstext, image | Digest7322db586141 | Download31 GB | Parameters27b | QuantizationMXFP8 | Context256K | Est. memory≈ 36 GB | ollama run clef:27b-mxfp8 |
| 27b-nvfp4 | Typelocal | Inputstext, image | Digestd1f6eb2a43fa | Download18 GB | Parameters27b | QuantizationNVFP4 | Context256K | Est. memory≈ 21.5 GB | ollama run clef:27b-nvfp4 |
| 27b-q4_K_M | Typelocal | Inputstext, image | Digest7885241a86db | Download20 GB | Parameters27b | QuantizationQ4_K_M | Context256K | Est. memory≈ 23.5 GB | ollama run clef:27b-q4_K_M |
| 27b-q8_0 | Typelocal | Inputstext, image | Digestb3d9f9fe2a55 | Download30 GB | Parameters27b | QuantizationQ8_0 | Context256K | Est. memory≈ 35 GB | ollama run clef:27b-q8_0 |
| cloud | Typecloud | Inputstext, image | Digest5896d5380b6c | DownloadUnknown | ParametersUnknown | QuantizationUnknown | Context64K | Est. memoryUnknown | ollama run clef:cloud |
| latest | Typelocal | Inputstext, image | Digestfdd5939e1bd9 | DownloadUnknown | ParametersUnknown | QuantizationUnknown | Context256K | Est. memoryUnknown | ollama run clef:latest |
Related models
devstral-small-2both
24B model that excels at using tools to explore codebases, editing multiple files and power software engineering agents.
qwen3.8local
Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks.
qwen3.6local
Qwen3.6 delivers substantial upgrades in agentic coding and thinking preservation than previous Qwen models.
muse-glimmerlocal
Meta's latest open model built for always-on local agents. 30B parameters, licensed under Apache 2.0 and runs on a single GPU — tuned for tool use, long tasks, and failure recovery.