AITOPAI热报
Products

Nace AI 推出预测市场决策模型 Drex for Finance

TestingCatalog@testingcatalog
Post
View the post
Why we picked it

做预测市场或概率判断类的开发者可以看看这套架构：Drex 自己打分，低于90%置信度才调 Claude 复核，108条全对还比纯用 Claude 便宜93%。分层调用的省钱思路很实用。

Nace AI 发布面向预测市场的决策模型 Drex for Finance。在 108 条测试预测中，Drex 全部答对，得分 100%，成本比 Claude 低 93%。

Body

工作流程上，Drex 会逐条读取预测市场的规则与最终价格，对每一方给出概率评分。当置信度低于 90% 时，系统会调用 Claude 进行二次复核，108 条测试中仅触发 7 次回退。

这种小模型为主、通用大模型兜底的混合架构，直接对标直接用 Claude 跑预测的场景，把高置信度判断交给专用模型可以显著压低成本。对做预测市场分析、量化判断类应用的开发者，这类分层调用思路值得参考。

Original · TestingCatalog

Nace AI has launched Drex for Finance, its decision model designed for prediction markets.

It reads each rule and the final price for the requested prediction, then gives a probability score for each side.

When Drex is less than 90% sure, it uses Claude to double-check the answer. This happened only 7 times out of 108. According to Nace, Drex scored 100% on all 108 predictions from the test suite, while being 93% cheaper than Claude.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Products" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Nace AI, Drex, Claude, reasoning models.

Source attribution: curated from TestingCatalog (@testingcatalog).

#Nace AI#Drex#Claude#reasoning models

Keep reading

Recent stories with the same tag or section
ProductsBuilders00:44
Chinese contentNace AI 发布 Drex 文档处理服务，修复扫描件比重新生成快 38%

Nace AI 推出的 Drex 能直接修复歪斜、模糊的扫描文档照片，官方测试显示比整页重新生成快 38%，每千页 7 美元，支持 PDF、表格、邮件等 30 种格式。经常处理扫描件和拍照文件的团队可以对比一下现有 OCR 方案的成本和速度。

#Nace AI#Drex#document processing#OCR
TestingCatalog@testingcatalogSource
ModelsBuildersMulti-source23:37
Chinese contentNace AI 发布 Drex Max v1.1，登顶 Artificial Analysis Decision Index

Nace AI 把 Drex Max v1.1 做到了 Decision Index 第一名，速度快还能处理多模态输入。做智能体或研究工作流的朋友可以直接调 API，有约 100 万次免费额度，开源权重也快放了，想在本地跑决策模型的可以关注。

#Drex Max#Nace AI#Decision Index#open-source models
Event
elvis@omarsar0Source
ModelsBuildersSelected02:50
Nace AI publishes 51-page report on Drex v1.5 training

Calibration training—tuning confidence to match real accuracy—is the detail that separates this from typical fine-tuning writeups, since miscalibrated probabilities are a common failure point for models used in decision support. Practitioners evaluating open-weight models for document-heavy workflows can compare Drex's approach against closed APIs and test it directly in Nace Console. The full training methodology is public, making it a useful reference for teams building constrained-answer systems.

#Nace AI#Drex#open-source models
TestingCatalog@testingcatalogSource
ModelsBuildersSelected00:56
Nace AI's 27B Drex Max Picks Instead of Generating Tokens

Routing thousands of micro-decisions to a pick-only model could cut agent orchestration costs by orders of magnitude where the answer space is a finite option list, not free text. Builders running multi-step agent pipelines can plug it in as a cheap decision layer while keeping a general model for generation. The gap on HLE shows it is a narrow tool, not a replacement for reasoning models.

#Drex Max#Nace AI#GPT-6#Claude Opus 5.5
Event
rohanpaul_ai@rohanpaul_aiSource