startlux-decision
startlux-decision is Typed decision models from 0.8B to 35B-A3B, over text and images with a 256K-token context. Every question comes back with a probability for each option. - StartLuxLabs/StartLux-Decision. It is ranked #523 on the Dev Radar, in Other, first seen 24 days ago and shared in 4 posts (19.3k views).
StartLux-Decision/docs/results.md at main · StartLuxLabs/StartLux-Decision · GitHub Skip to content Navigation Menu Sign in Appearance settings Search / Sign in Sign up Appearance settings You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert {{ message }} Repository navigation Files Expand file tree main / results.md Copy path Blame More file actions Blame More file actions Latest commit History History History 412 lines (329 loc) · 24.5 KB main /…
What people said about startlux-decision on X
🤯无敌了！0.8B~27B全尺寸开源！全球最强决策模型StartLux-Decision炸场，算力只花在刀刃上！ StartLux Labs刚发布的StartLux-Decision（覆盖0.8B、2B、4B、9B、27B）彻底改写决策模型赛道： 🔹Instant：一次前向直接出答案，单卡H200中位延迟仅25-52ms，快到离谱。 🔹AutoThink：有把握就秒答，没把握才短推理，智能分配算力。 🔹原创RLCD²：用冻结模型+Proper Scoring Rule打分，奖励“概率变准”，软性惩罚过长推理 → 平均推理长度最多缩65%，长度上限内完成率最多提48个百分点，准确率几乎不掉。 🔹Decision…
— @NFT_Chen, 13 days ago · 151 likes · see the post
💥太牛了！Agent交互强化学习终于不再“工程地狱”了！HarborRL开源全套框架，真实Coding Agent也能稳稳做RL！ 以前训Agent RL最痛苦的不是算法，是轨迹收集、Token对齐、多轮工具调用、策略版本一致性这些工程坑。HarborRL把它们全部打通，做成可复现流水线。 核心亮点 🔹直接复用Harbor沙箱任务（200+，含Terminal-Bench/SWE-Bench等），自带Verifier算Rewar 🔹支持真实Coding Agent多轮Rollout（Claude Code等40+ harness）+ SGLang推理 🔹轨迹严格校验 + 无缝接入Slime/Megatron训练…
— @NFT_Chen, 1 days ago · 62 likes · see the post
💥我靠！LLM评估终于不用再靠另一个“爱飘”的生成式裁判了！ DeepEval正式推出 JevEval：用Jev当自定义评估裁判，而不是每次都让另一个LLM去“生成”评分。 对比传统LLM-as-a-Judge： 🔹传统：模型自己解读标准 + 推理 + 生成结果（易飘、贵、慢、JSON还容易炸） 🔹JevEval：决策结构你定好，权重你配，strict_mode一开就是硬门槛。支持Yes/No、分级Score、Choice三种题型 特别适合Agent：直接评执行轨迹（trace），工具选对没、动作顺序对不对、任务到底完成没。 关键操作步骤（3分钟上手）： 1️⃣安装依赖：pip install -U deepeval…
— @NFT_Chen, 6 days ago · 42 likes · see the post
Alternatives to startlux-decision
- Step-5-Preview — Join the countdown
- Google — Argon is rolling out today to Google Ultra and Pro subscribers in the @GeminiApp and Search, and Ultra subscribers in…
- nanbeige4.3-33b — We’re on a journey to advance and democratize artificial intelligence through open source and open science.
- Spark-X2.5 — Healthcare-focused AI just got a little lighter. Same powerful reasoning, less memory to worry about. 🤖
- anymove — Open source
- Ideogram — Live in Ideogram and the API
startlux-decision in numbers
- Rank on the Dev Radar: #523 of 8801
- Shared in 4 posts by 3 accounts: @NFT_Chen, @Angelinafox_AI
- 19.3k views on those posts
- First seen 24 days ago, last shared 1 days ago
- Pricing seen by AGK Intelligence: open source
- Market: Other
FAQ
What is startlux-decision?
Typed decision models from 0.8B to 35B-A3B, over text and images with a 256K-token context. Every question comes back with a probability for each option. - StartLuxLabs/StartLux-Decision It was first shared on X 24 days ago and is ranked #523 on the Dev Radar.
Is startlux-decision free?
It is open source.
Who shared startlux-decision?
3 accounts on X, including @NFT_Chen, @Angelinafox_AI, in 4 posts totalling 19.3k views.
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.2k posts from 5k X accounts over the last 21 days, 8.8k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-10-30 21:12 UTC. Full method.