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
Claude Fable 5 (high)
6 atoms · 跨 2 天 · 首见 2026-06-09 · 最近 2026-06-11
三色: 🟦 fact 5 · 🟥 take 1 · stance ▲3/▼1/◆0
来源: X 6
时态: fresh:6
标签: 好数字:4 · 好信源:3 · 好观点:1
叙事 (narrative) (1)
- 2026-06-11 · inference cost perception
waow, not even that expensive great!!
tags: 好观点 · → daily
事实 (fact) (5)
- 2026-06-11 · WeirdML score
Claude Fable 5 (high) scores 87.8% and takes the lead on WeirdML.
tags: 好数字·好信源 · → daily
- 2026-06-11 · output token usage
It uses about 8k output tokens on average
tags: 好数字·好信源 · → daily
- 2026-06-09 · short-fiction writing 能力提升
Claude Fable 5 (high) is a step up in short-fiction writing.
tags: 好信源 · → daily
- 2026-06-09 · Short-Story Creative Writing Benchmark rank
On the Short-Story Creative Writing Benchmark, it beats Claude Opus 4.8 (xhigh) and Claude Opus 4.7 (high), and ranks second behind GPT-5.5 (xhigh).
tags: 好数字 · → daily
value: qty=second behind GPT-5.5 (xhigh)
- 2026-06-09 · 谨慎拒绝率
Caveat: it refused 5 of the 400 creative-writing prompts.
tags: 好数字 · → daily
value: qty=5 out of 400 prompts
时间轴 (近 20)
- 2026-06-11 ·
fact · WeirdML score · → daily
- 2026-06-11 ·
fact · output token usage · → daily
- 2026-06-11 ·
narrative · inference cost perception · → daily
- 2026-06-09 ·
fact · short-fiction writing 能力提升 · → daily
- 2026-06-09 ·
fact · Short-Story Creative Writing Benchmark rank · → daily
- 2026-06-09 ·
fact · 谨慎拒绝率 · → daily
← 实体目录 · 系统日志