openballast

Quantize the weights. Ballast the knowledge.

Small models reason fine — they just don't know much. Parameters are the most expensive place to store facts: a fact costs ~100× more bytes in weights than in a compressed triple corpus. Ballast is that corpus: a versioned, CC0, rank-bucketed knowledge artifact you pair with any local model, quantized in levels like a GGUF.

+180 MB > +19 GBcorpus bytes lifting Gemma-4-E2B past a 12B's factual accuracy (measured, PopQA+4 sets)

3× less hallucinationE2B confident-wrong rate 0.24 → 0.07 with full ballast

$0 / monththis entire knowledge layer — 62.6M triples — runs on Cloudflare's free tier

Try it now (demo endpoint)

Levels L0–L5 served live from a $0/month free-tier stack — demo-grade, no auth, no SLA. Point any MCP client at:

https://mcp.openballast.org/mcp

Or plain HTTP:

curl "https://mcp.openballast.org/lookup?question=Where+was+Douglas+Adams+born%3F&level=5"
curl "https://mcp.openballast.org/evidence?id=Q42&level=3"

Get the artifact

Canonical downloads live on Hugging Face — hive-partitioned parquet, 8 nested quantization levels (36 MB → 1.51 GB), CC0:

huggingface.co/datasets/OpenBallast/ballast-t0

Eval sets (50k linked probes): OpenBallast/ballast-evalsets · Thesis, methodology, numbers: github.com/OpenBallast/ballast

What it is

T0 = 197M cleaned Wikidata triples (CC0), entity-ranked by a composite of sitelinks and claim count, published as nested rank buckets. Truncation is the quality knob: drop deep buckets and evidence thins gracefully — measured rate–distortion, not vibes. A model-aware profile then selects tuned ballast: only what your model doesn't already know.

Data: Wikidata, CC0. Artifacts on Hugging Face; thesis and docs on GitHub. Contact: kevin@openballast.org