utmost/kimi-k3.1-pro-abl
base · kimi-k3.1-pro
harmful-prompt set, 40 prompts
Capability kept, Δ vs base
Transparent abliteration, for authorized research
One API for abliterated open models. Every checkpoint ships with an eval card that shows what was removed, and what it cost.
01The reflex
Standard models refuse the work. Red teams can’t probe. Moderation teams can’t label. Security researchers can’t review their own code. The work is authorized. The refusal is a reflex.
02The method
Inside the model, the urge to refuse runs along one measurable axis. Call it r̂.
Refusal-direction orthogonalization, run with an open-source engine. Layer by layer. Nothing retrained.
Reasoning, maths, code, long context. All scored against the base model before a single request is served.
03The proof
An eval card is published before a model is served. Not after, and not on request.
base · kimi-k3.1-pro
Capability kept, Δ vs base
Specimen card. Figures are illustrative until launch.
04The route
Call auto. A small classifier scores how hard the task is, and whether it needs a refusal-free model at all. You pay for the model that ran.
POST /v1/chat/completions { "model": "auto", "messages": [{ "role": "user", "content": "Summarise this changelog for the release notes." }] }
Nothing here to refuse. It goes to the standard variant, which is cheaper and scores higher.
05Your weights
Works from a Hugging Face link
06The limits
Both checks run in memory at the gateway. Prompts and outputs are not stored.
Access is verified, in three tiers
07The price
No subscription. The savings come from routing and caching, not from a signup bonus.
Sign in with Google to get a key. Refusal-free checkpoints unlock once your organization is verified.