跳到正文
RadarAI· X:Rohan Paul (@rohanpaul_ai)·· 3 小时前AI 评分61

MachinePulseAI 开源 Karpo Decision 系列决策模型,31B 版本以 73.9 分并列 Decision Index 榜首

开源决策模型 Karpo Decision 31B 发布,以 73.9 分并列 Decision Index 300 模型榜首

AI 导读

MachinePulseAI 发布开源权重决策模型 Karpo Decision 系列,含 0.8B 到 31B 六个型号,用单次前向传播直接输出各选项概率而不生成 token。

正文 · 原文

Rohan Paul· @rohanpaul_ai · X·2026-11-18 03:14· 58 分钟前AI 评分50

AI 导读

MachinePulseAI 发布开源权重决策模型 Karpo Decision 系列,含 0.8B 到 31B 六个型号,用单次前向传播直接输出各选项概率而不生成 token。

正文 · 原文

A new open source decision model just launched, and it ranks first out of 300.

Karpo-Decision-31B swaps decoding for a readout head: a single-token code per option, one score per code, a softmax over that question's options. Zero generated tokens, always a valid answer, and the page is shared across questions through branch attention.

Decision Index Full: 73.9, the top score, tied #1 of 300 with GLiDE 27B at 73.4. The base is Gemma 4 31B, yet it beats Google's Gemini Decision 1 on Amazon ESCI, 64.7% vs 58.7%, and New Yorker Caption, 90.5% vs 79.7%.

Against Jev 1.14 it runs 95 ms vs 196 ms median, and JevBench estimates about a fifth less per 1,000 decisions.

Open weights matter for search and recommendation teams, who need many relevance calls per query, and for personal agents that need a calibrated act-or-ask threshold.

引用LeahW@LeahW_2077

Picking the next tool shouldn’t take a paragraph.

We built Karpo Decision. A family of six open-weight models, from 0.8B to 31B.Give it the context, a question, and the options. It returns a probability for each option in one forward pass.

Karpo Decision 31B
🏆 73.9 on Decision Index, tied for #1 on the Nov 12 board
⚡ 94 ms median per request on a single RTX PRO 6000
📊 87.2 on JevBench Capability vs. 78.5 for Google’s Gemini Decision 1

🤗 Hugging Face
huggingface.co/MachinePulseAI…
📝 Technical write-up
karpo.ai/blog/karpo-decision-…

Here’s where I’d try it first: Search. Recommendations. Tool choice. Model routing. Browser actions. Point it at the slowest decision in your stack and tell us what breaks.

在 X 查看被引用的帖子

来源:Rohan Paul · x.com

来源:RadarAI · aihot.news