以下部分引用 AI 总结,现阶段 AI 仍然有幻觉。请以内容的原文为准。
数据新鲜度群聊 07-14 ✓卖方 07-14 ✓Wrap 断档15天 ✕News 07-15 ✓X 断档12天 ✕页面生成 07-15
Mythos/Fable
5 atoms · 跨 4 天 · 首见 2026-06-09 · 最近 2026-06-17
三色: 🟦 fact 3 · 🟥 take 2 · stance ▲2/▼1/◆1
来源: X 5
时态: fresh:5
标签: 好信源:3 · 好思考:2 · 好观点:1
叙事 (narrative) (2)
- 2026-06-09 · agentic coding eval performance
Mythos/Fable destroys everything else in everything all at once, with large gaps. On CritPt, they are merely equal with 5.5 (and below 5.5-Pro.)
tags: 好思考 · → daily
- 2026-06-09 · 被描述为最智能的模型
mythos/fable is by far the smartest model ever released
tags: 好观点 · → daily
value: direction=正面 · date=2026-06-09
预测 (forecast) (1)
- 2026-06-17 ⭐ · 监管政策改变影响范围
a de facto policy change that limits not just Mythos/Fable but all future models to administration-approved entities will destroy the economics of frontier models
tags: 好思考·好信源 · → daily
事实 (fact) (2)
- 2026-06-13 · release tiered as Glasswing first to cybersec firms and NSA, then banks, then 150 orgs, then public as Fable 5
Mythos/Fable was a tiered release. It started as Project Glasswing where cybersec firms & NSA got access first, then it spread to banks and finally it was expanded to 150 orgs globally. Then they did a public release as Fable 5
tags: 好信源 · → daily
- 2026-06-12 · 能运行在 Google TPUs 或 Trainium 上
Yes, it does in prod
tags: 好信源 · → daily
value: qty=yes · date=prod
时间轴 (近 20)
- 2026-06-17 ·
forecast · 监管政策改变影响范围 · → daily
- 2026-06-13 ·
fact · release tiered as Glasswing first to cybersec firms and NSA, · → daily
- 2026-06-12 ·
fact · 能运行在 Google TPUs 或 Trainium 上 · → daily
- 2026-06-09 ·
narrative · 被描述为最智能的模型 · → daily
- 2026-06-09 ·
narrative · agentic coding eval performance · → daily
← 实体目录 · 系统日志