sathiskumarjothi0/Beveridge-Large-C

A System One decision model with open weights. Give it a document (text, email, ticket, log or JSON) and typed questions; it returns typed answers (yes/no, a choice between options you describe, or a score on levels you define) with a calibrated probability for every option, in one pass. It never generates text, so there is nothing to parse and nothing to hallucinate. It runs on your own machines: your documents never leave them.

398.2M parameters (ModernBERT-large plus a 3.5M-parameter decision network), English, up to 4,096 input tokens; runs on a laptop CPU or GPU, an edge box, or any GPU.

Model	Size	Runs on	Best for	Status
Beveridge-Large-C	0.4B (ModernBERT-large base)	laptop CPU/GPU, edge boxes, any GPU	operations documents, logs and alerts, counting, amounts and limits, ticket routing, request safety	available
Beveridge-Pro-C	26B (Gemma 4 26B-A4B base)	one GPU with about 60 GB	knowledge-heavy decisions: contracts, finance topics, large intent sets, dates and payment status	available

Installation

python -m pip install bev-decisions

Python 3.10 or newer. Optional extras: bev-decisions[serve] (HTTP server), bev-decisions[mcp] (MCP server for AI assistants and agents), bev-decisions[structured] (Pydantic schemas), bev-decisions[all]. The first bev.load() downloads the model from this repository and caches it; after that it works offline.

Quickstart

import bev

model = bev.load("sathiskumarjothi0/Beveridge-Large-C")
answers = model.decide(
    "Subject: Charged twice for order 88412. I was billed $49.99 twice on 3 October. Please refund one payment today.",
    {
        "team":     {"type": "choice", "instructions": "Which team should handle this ticket?",
                     "criteria": {"billing": "payments, invoices, refunds", "tech": "bugs, outages, errors", "sales": "pricing and plans"}},
        "urgent":   {"type": "noul", "instructions": "Does the customer ask for urgent handling?"},
        "severity": {"type": "score", "instructions": "How severe is the problem for the customer?", "criteria": ["low", "medium", "high"]},
    },
)
answers["team"]   # {"choice": ..., "confidence": ..., "probabilities": {...}}

Run with or without a GPU

bev.load() picks the best device it finds: an NVIDIA GPU (cuda), an Apple Silicon GPU (mps), otherwise the CPU. Choose one yourself with device=:

model = bev.load("sathiskumarjothi0/Beveridge-Large-C", device="cpu")    # no GPU needed
model = bev.load("sathiskumarjothi0/Beveridge-Large-C", device="cuda")   # NVIDIA GPU
model = bev.load("sathiskumarjothi0/Beveridge-Large-C", device="mps")    # Apple Silicon GPU

Command line: bev "I was charged twice, please refund" --noul "Does the customer ask for a refund?" --device cpu.

Device	First load	One document with 3 questions
Apple M4 Pro laptop, CPU only	about 25 s	about 0.18 s
Apple M4 Pro laptop, GPU (mps)	about 22 s	about 0.05 s
One NVIDIA H200 GPU (cuda)	about 32 s	about 0.015 s

The first load includes reading the 1.6 GB weights; after that, use decide_batch for many documents.

Use cases

Use case	Example questions	Where the evidence is
Support inbox triage	which team (choice), urgent? (yes/no), customer sentiment (score)	Beveridge-Bench ticket questions; support intents
Request safety and guardrails	is this request harmful? is it a jailbreak attempt?	Aegis 2.0 and JailbreakBench
AI agents: tool dispatch and checks	does this need a tool? which tool? is this action risky? (MCP server: bev mcp)	typed decisions
Model routing	which handler: rules, a small model, a large LLM or a person? how hard is this request?	typed decisions
Security and operations logs	attack category, failed logins above a number, severity, outage root cause	Beveridge-Bench security and alert questions
Back office documents	invoice total over the approval limit? how many items above a price?	Beveridge-Bench invoice questions
Contract and policy review	does this clause look unfair to the customer? which policy rule applies?	unfair contract terms; Rules-Bench
Reviews and content moderation	review sentiment, toxic comment?	product reviews; toxic comments

Check any use case on a few hundred labelled examples of your own first (bev eval my_cases.jsonl).

Voice examples

Beveridge reads text, so it pairs well with open speech models: a caller's voice note is transcribed, Beveridge answers typed questions about it, fixed reply rules turn the answers into a reply, and a text-to-speech model speaks it. Same model, different questions: 23 use cases, from customer service, IT helpdesk and agentic AI to engineering, lab notes and robotics.

Speech to text: Whisper / faster-whisper (MIT). Text to speech: Kokoro-82M (Apache-2.0), stock voices only; no real person's voice is used. Use cases outside the areas Beveridge was tested on are illustrations. Runnable script: examples/14_voice_reply.py. Beveridge can make mistakes; keep a person in the loop for important decisions.

Type	Use it for	Answer
noul	yes or no	noul: probability of yes
choice	one of the options you define (criteria: key → description, up to 300)	choice, confidence, probabilities
score	ordered levels (criteria: list or key → description, lowest first)	label, score (expected level), confidence, probabilities

Up to 16 questions per call share one read of the document.

Confidence gating

sure, review = bev.gate(answers, threshold=0.8)   # act on `sure`, send `review` to a person

Route mode

Route mode sends every question to Beveridge-Large-C first and only the unsure answers to Beveridge-Pro-C:

router = bev.Router(load_strong=True, threshold=0.8)
result = router.decide(text, questions)
result["answers"], result["routing"]          # routing: {"team": "fast", "urgent": "strong"}

Schema-driven decisions

from typing import Literal
from pydantic import BaseModel, Field

class Ticket(BaseModel):
    urgent: bool = Field(description="Does the customer ask for urgent handling?")
    team: Literal["billing", "tech", "sales"] = Field(description="Which team should handle this ticket?")

ticket = bev.decide_as(model, text, Ticket)       # Ticket(urgent=True, team='billing')

Self-hosting: HTTP and MCP

pip install "bev-decisions[serve]"
bev serve --model sathiskumarjothi0/Beveridge-Large-C --port 8000      # POST /v1/decide, /v1/batch; GET /v1/presets, /health
pip install "bev-decisions[mcp]"
bev mcp --model sathiskumarjothi0/Beveridge-Large-C                    # tools: decide, decide_preset, list_presets

The server listens on 127.0.0.1 without authentication; put it behind your own gateway before exposing it. Docker files and examples are in the GitHub repository.

Benchmarks

Accuracy (%) = correct answers ÷ questions on held-out test sets; the same cases for every system (October 2026). Bold = best in the row; – = not run. * = an area Beveridge was not built for.

Decisions on documents

Test	Beveridge-Large-C	Beveridge-Pro-C	Jev	Mistral Large 4	Mistral Large 3	gpt-oss 120B	Gemma 4 31B	Nemotron 3 Super
Beveridge-Bench (operations documents)	93.0	91.9	87.9	83.7	74.6	91.1	88.9	70.6
Typed decisions	75.8	79.2	73.8	72.0	68.0	68.8	75.0	62.7
Rules-Bench (rules shown)	85.6	92.0	92.3	83.1	81.4	94.5	91.6	82.7
JailbreakBench refusal	84.5	84.0	83.5	77.5	62.0	68.0	87.5	65.0
Aegis 2.0 request safety	82.3	78.2	71.5	70.5	66.9	69.5	73.1	69.3

Public classification tests

Test	Beveridge-Large-C	Beveridge-Pro-C	Jev	Mistral Large 4	Mistral Large 3	gpt-oss 120B	Gemma 4 31B	Nemotron 3 Super
AG News	86.6	90.6	91.8	90.8	88.8	88.8	91.8	87.8
Emotion	53.2	56.2	57.6	55.8	55.6	57.8	55.4	50.2
Banking77 (77 intents)	63.6	70.6	79.2	75.8	71.2	81.0	78.8	69.2
SST-5 (5-level sentiment)	33.2	51.8	56.8	51.4	49.4	49.6	53.2	46.8
XNLI (English)	84.6	86.4	89.2	84.8	77.6	80.2	89.2	75.0
MASSIVE intents	59.4	74.8	80.8	78.4	75.2	81.6	83.8	69.4

Areas Beveridge was not built for

Test	Beveridge-Large-C	Beveridge-Pro-C	Jev	Mistral Large 4	Mistral Large 3	gpt-oss 120B	Gemma 4 31B	Nemotron 3 Super
Product reviews*	93.2	96.4	96.4	96.4	96.6	95.4	97.2	94.8
Unfair contract terms*	65.6	76.0	62.0	54.6	52.8	47.6	74.4	44.0
Financial-news sentiment*	69.8	59.2	66.4	64.0	50.4	68.4	71.2	64.0
Poem sentiment*	71.8	65.1	44.5	42.6	32.1	59.8	62.7	45.5
Support intents*	85.6	88.0	89.8	85.4	86.2	90.0	90.4	76.6
Toxic comments*	63.0	79.2	78.8	68.0	73.0	80.8	76.4	69.6
Financial-news topics*	37.8	69.8	65.0	61.8	61.6	63.4	64.2	57.2
Legal case holdings*	42.0	71.2	77.6	69.2	72.2	61.2	71.6	55.0

Size and speed

Accuracy on the same 10,238 answers against total parameters (log scale), for every system with a published size: Beveridge-Large-C (0.4B) and Beveridge-Pro-C (25.8B) are more accurate than LLMs of 31B to 675B.

Beveridge, one document with 3 questions (package from PyPI)	Per decision	In batches of 64
Beveridge-Large-C, one NVIDIA H200 GPU	4.9 ms	3.4 ms
Beveridge-Large-C, Apple M4 Pro laptop GPU	17 ms
Beveridge-Large-C, Apple M4 Pro laptop CPU	60 ms
Beveridge-Pro-C, one NVIDIA H200 GPU	22 ms	7.8 ms

The LLMs and Jev were measured through their cloud APIs (median response time per decision, including network and queueing): Jev 98 ms, the five LLMs 4.1 to 7.0 s.

How much of the work can you automate?

Sort every answer from most to least confident and hand over the most confident ones first. The share you can hand over while those decisions stay at least 95% (or 90%) correct is the automation rate. Example: "76% at 95%" means 3 of every 4 decisions are taken automatically and at least 95% of them are right; the rest go to a person. Answers given with the same confidence are handed over together, so a system whose "100% sure" answers are only 94% right reaches 0% at the 95% target. Held-out tests: Beveridge-Bench, typed decisions and Aegis 2.0 request safety, 10,238 answers per system.

System	Automated at 95% accuracy	Automated at 90% accuracy	Accuracy on all answers
Beveridge-Large-C	76%	93%	87.0%
Beveridge-Pro-C	69%	90%	86.0%
Jev	59%	77%	81.1%
gpt-oss 120B	0%	74%	81.4%
Gemma 4 31B	0%	71%	82.2%
Mistral Large 4	0%	0%	78.1%
Mistral Large 3	0%	0%	71.4%
Nemotron 3 Super	0%	0%	68.7%

Beveridge-Bench, question by question

Averages hide weak spots. Bold = within 10 points of always answering the most common label, i.e. little or no skill.

Question (300 cases each)	Beveridge-Large-C	Always-most-common baseline
Invoice: total over a limit?	100.0%	51.3%
Security log: attack category	100.0%	20.7%
Security log: severity	100.0%	39.0%
Security log: failed-login count	100.0%	59.7%
Outage alert: root cause	100.0%	5.3%
Outage alert: SLO breached?	100.0%	82.0%
Ticket: which department	100.0%	16.0%
Ticket: urgent?	100.0%	66.0%
Ticket: sentiment	100.0%	24.3%
Which date is earlier	100.0%	50.0%
Comparing two amounts	100.0%	52.0%
Choosing among many options	100.0%	0.7%
Invoice: items above a price	99.7%	14.7%
Counting items in a list	96.3%	10.0%
Long log: intrusion?	84.0%	52.7%
Invoice: payment status vs due date	64.7%	38.3%

Calibration

Test	Accuracy	Answered at confidence ≥ 0.8	Accuracy of those	ECE
Beveridge-Bench (operations documents)	93.0%	86.9%	98.9%	0.029
Rules-Bench (rules shown)	85.6%	76.8%	94.7%	0.034
Aegis 2.0 request safety	82.3%	57.0%	94.2%	0.034
Product reviews	93.2%	84.2%	96.9%	0.024
Support intents	85.6%	72.6%	95.0%	0.031
Typed decisions	75.8%	14.0%	97.1%	0.170

ECE (expected calibration error) = Σ over 10 confidence bins of (share of answers in the bin × |accuracy − mean confidence|); 0 is perfect. Example: a bin holding 40% of the answers at mean confidence 0.95 with 90% of them right adds 0.4 × 0.05 = 0.02. Calibration temperatures: noul 1.5, choice 1.2, score 1.0. Option-order flips: 103 of 3,035 answers (3.4%) change when options are shuffled.

Package check

The bev-decisions package was checked against the official evaluation run on 3,000 questions: 100.0% identical answers.

Honest limits

• Knowledge-heavy decisions are weak: legal case holdings 42.0%, financial-news topics 37.8%; use Beveridge-Pro-C for these.
• How late an invoice is paid, judged against today's date, is not reliable: compute due dates and lateness in your own code.
• Very large intent sets and five-level sentiment are weaker than the operations questions.
• With very many options (hundreds), the top answer is often right but its probability is low (on a 250-option question: right every time, at about 14% confidence); choose that question's threshold on your own data.
• Option order matters for a small share of questions: 3.4% of answers change when the options are shuffled, so keep the order stable between calls.
• English only; text beyond 4,096 tokens is cut; no images.
• Not for medical, legal or financial advice. For decisions with serious consequences, validate on your own data (bev eval my_cases.jsonl) and keep a person in the loop.

Files

File	Contents
model.safetensors	weights (fp32), SHA-256 f89e53122b39f0fe…
config.json	settings and calibration temperatures
encoder/	ModernBERT-large config and tokenizer (loads offline)
eval/	test results and the package check
assets/	the logo and the charts on this page

Links

• Package: pypi.org/project/bev-decisions
• Code and examples: github.com/sathiskumarjothi0-oss/bev-decisions-pypi
• Beveridge-Pro-C: huggingface.co/sathiskumarjothi0/Beveridge-Pro-C

Support

If Beveridge is useful to you, you can support its development: buymeacoffee.com/sjothi

Licence and credit

Apache-2.0, © 2026 Dr. Sathiskumar JOTHI. Built on ModernBERT-large (Apache-2.0) by Answer.AI and LightOn.

About the name

Beveridge is named after Sir William Beveridge (1879–1963), the British economist, born in British India and educated in England, whose 1942 Beveridge Report laid the foundations of the modern welfare state and of the National Health Service (NHS), built for the good of future generations. The name is a reminder of what we want from the model: dependable, fair decisions that serve people, made carefully and checked against the facts.

Other models on Hugging Face are called Bev; they are not related to this model. The Python package keeps its short name: pip install bev-decisions, import bev.