Jeff-Qwen3.5-0.8B

Use it in local code to make fast classification decisions, such as routing support requests, detecting spam, or choosing a tool.

Availability

Community preview: v1.2 (1 October 2026); v1.1 remains available by revision

Pricing model

Not established

Deployment

Built and tested on local hardware; no cloud GPUs used for base-model training

Privacy

No cloud GPUs used for base-model training

Licence

Licences differ by adapter; sanctions and soc are CC BY-NC 4.0, while aml and trading-desk have data-source terms.

• The v1.3 base and adapters are available as Hugging Face downloads, with GGUF versions also provided. ^[1]

  Weights: `mstrasser/jeff-adapter-<name>` on Hugging Face (revision `v1.3`), with the GGUF versions in

• The quick start supports NVIDIA GPUs and Apple silicon. ^[1]
• The announcement lists 15 adapters for task-specific decisions. ^[1]

  for name in guard tools ground; do # the adapters you want, one folder each

• The source recommends v1.2 for zero-shot use without an adapter. ^[1]

  For zero-shot use without an adapter, use v1.2.

• The page describes v1.2 as a community preview and links to its Hugging Face model repository.
• The model returns a calibrated probability for each option from a single forward pass. ^[2]

  probability for each option from a single forward pass.

• The Qwen models support choice questions with up to 254 options. ^[2]

  question types: `choice` (pick one of up to 254 options with the Qwen models, 26 with Jeff-Gemma4-E2B), `noul` (yes/no, returned as a probability) and `score` (a

• The v1.2 benchmark-panel score is 78.7% across five benchmarks. ^[2]
• The model takes about 22 ms per decision on an RTX PRO 6000, according to the page. ^[2]

  **Built entirely on local hardware.** Training on one RTX PRO 6000 workstation GPU (the 0.8B trains in about 2 hours,

• Nine LoRA adapters are available for v1.2 and can be selected by name in a request. ^[2]

  v1.2 is also the base for nine small LoRA adapters (legal clauses, support

• OpenAI introduces GPT-6.1 Sol for coding and multi-step workflows Daily brief 2026-09-29