Six demos, from games to guardrails.
Agentic pipelines5 min read
Quirq’s XO platform used Sage to select models, then retried unfinished work on a stronger model. Across 90 tasks, it reached similar final completion for $34.04 instead of $69.57.
51% Lower evaluation cost
Read
Gaming
An open-source harness gave a dwarf priest a decision model. 4,573 decisions later it hit level five, at under three cents per kill.
$3.32 Sage cost, level 1 to 5
View
Finance4 min read
Claude Code delegates thousands of market classifications to Sage, keeping bulk results in files while Python tests the findings against predefined criteria.
28× Faster research workflow Customer-reported workflow timing and research results
Read
Education3 min read
Leap routes requests for flashcards, quizzes, and other study activities from chat into native tools, with 89.9% accuracy in its reported evaluation.
89.9% Tool selection accuracy Leap benchmark · 327 held-out cases, each run three times
Read
Gaming3 min read
Two authoring passes. 40 monster types. GrowBlocks brings natural-language personalities into combat decisions and adaptive dungeon encounters.
40 Monster types Inside Rogblox · reported by the project team
Read
Finance2 min read
Six structured decisions help Capacitr filter market noise, focus research, and check its trading agent before expensive work begins.
~44% Lower pipeline LLM spend Projected monthly impact
ReadBuild with Sage