Our models
relia-1
Your relevance model for financial context.
Give it an entity and a batch of articles — relia-1 scores how relevant each one is, so you filter the noise before it ever reaches your pipeline, dashboard, or LLM.
Efficiency
Exact-bin agreement with a 3-LLM consensus
ROI
To score 1,000 articles — up to 500× cheaper than LLM APIs
Quality
Of the noise cut before it ever reaches your pipeline
Under the hood
What Relia returns
Give it an entity and a batch of articles — relia-1 returns a 0–1 relevance score and a tier for each, ranked so you can threshold or filter the noise in one pass.
Sample output
relia-1Tesla Q3 deliveries beat Wall Street estimates
High0.87Elon Musk unveils next-gen Optimus robot at demo event
High0.84New federal EV tax-credit rules could reshape automaker incentives
Low0.31Federal Reserve holds interest rates steady amid cooling inflation
Irrelevant0.13The problem
News feeds are flooded with articles that only mention a ticker in passing. Deciding what’s actually material is slow by hand and expensive with LLMs.
Why relia-1
Purpose-trained for finance, deterministic, no prompt engineering. One flat price per request of up to 100 articles — long articles are never re-chunked.
Benchmarks
Relia validation, measured.
Benchmarked against a 3-LLM consensus on accuracy, speed and cost.
Matches frontier LLMs
Exact-bin agreement with a 3-LLM consensus on a held-out test.
One request per batch
Up to 100 articles scored in a single pass — no per-article round-trips.
Up to 500× cheaper
Score a thousand articles for a fraction of an LLM API bill.