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
Moonshot
47 atoms · 跨 28 天 · 首见 2026-04-07 · 最近 2026-06-30
三色: 🟦 fact 22 · 🟥 take 25 · stance ▲16/▼8/◆10
叙事状态: 🍂 fading (退潮) · 💤 沉寂 · 策展近14d 0 atom · 🐦 X 近14d 2
状态只算低频策展源 (群/卖方); X firehose 仅作背景音量
来源: X 47
时态: fresh:39 · aging:6 · stale:2
标签: 好观点:21 · 好信源:16 · 好数字:11 · 好思考:4
🟨 AI 综合 · junior analyst 概览
展开 AI 综合 (灰色 · 非市场结论 · 点击数字溯源到原 atom)
model: deepseek-chat · 2026-07-12 · 默认折叠
Moonshot 这家中国AI初创刚完成美团的20亿美元融资,估值冲到300亿美元
¹¹¹¹。在算力上他们既被描述为“相对贫算力”,又能靠弹药拿到GW级算力
¹¹。产品端分歧明显:一边说Kimi K2.6过度思考、慢得离谱
¹¹,另一边称Moonshot的推理智能体始终领先DeepSeek 4-6个月
¹,而且即将发布的七月版本比DeepSeek完整版更值得期待
¹。开源策略上M3模型确定不会开源
¹,战略窗口也面临被阿里和字节挤死的风险
¹。
🧭 拥挤度 (一人一票): ▲ 8 位作者 (KOL8) vs ▼ 3 位作者 (KOL3)
⚖️ 多头 8/8 来自KOL
📄 广泛报道的事实 · 1 个
展开 (多人转述同一事实/数字 — 确认度高, **非独立观点**, 不标多空)
💢 核心分歧 (2 轴)
1. Moonshot 竞争窗口 紧逼 vs 护城河
*topic: 技术路线/产品发布/模型发布/竞争格局/模型性能 · 14 bull vs 7 bear*
🟢 bullish 侧:
- <a id="atom-29647764f0a232e1"></a>2026-04-26
X·@teortaxesTex [老化中] · K2.6 model strength relative to DeepSeek V4 · 技术路线 +同日1条
I think K2.6 is stronger than V4 though.
tags: 好观点 · → daily
value: direction=stronger
- <a id="atom-04c47284713c121e"></a>2026-05-01
X·@teortaxesTex [老化中] · 通过强化学习可在推理方面超越DeepSeek · 竞争格局/技术路线
for now Moonshot or w/e can outpace them with muh RL
tags: 好观点 · → daily
- <a id="atom-0fe1ab65d3a63d44"></a>2026-05-28
X·@teortaxesTex [老旧] · 推理智能体开发进度对比DeepSeek · 技术路线/竞争格局
They're consistently 4-6 months ahead of DeepSeek on the agentic stuff.
tags: 好观点 · → daily
value: qty=4-6个月 · direction=ahead
- <a id="atom-48836216b5dfa2b5"></a>2026-06-25
X·@Presidentlin · shipping almost SOTA models · 技术路线/产品发布
Moonshot still shipping almost SOTA models
tags: 好观点 · → daily
- <a id="atom-37a406f7424cb5fd"></a>2026-06-30
X·@tphuang · GLM-5.2 在 ZCode 上体验优于 DeepSeek V4 和 Kimi Code K2.7 · 模型发布/竞争格局 +同日1条
if you try GLM-5.2 on ZCode, it's an entirely different experience than V4 or even K2.7 with Kimi Code (different in a good way)
tags: 好观点·好思考 · → daily
🔴 bearish 侧:
- <a id="atom-62399a513b510678"></a>2026-04-18
X·@zephyr_z9 · M3 model open source status · 技术路线 +同日1条
M3 won't get open sourced
tags: 好观点 · → daily
- <a id="atom-f6dc82dc9b2180b9"></a>2026-04-07
X·@bioshok3 · 被Anthropic识别为蒸馏行为方 · 竞争格局/技术路线
今年2月にはDeepSeek・Moonshot・MiniMaxの3社を蒸留行為で特定
tags: 好信源 · → daily
value: date=2025年2月
- <a id="atom-4ba3e96389458781"></a>2026-05-24
X·@thegenioo · Kimi K2.6 表现评价 · 产品发布 +同日2条
Honestly I think Moonshot fumbled big time with Kimi K2.6
tags: 好观点 · → daily
- <a id="atom-96cda960c60d80a9"></a>2026-04-28
X·@teortaxesTex [老化中] · 战略窗口期 · 竞争格局
there is a small window now for Minimax/Moonshot/Zhipu to make excellent moves and secure their stake in the future, otherwise they're getting squeezed to death between Alibaba&ByteDance and DeepSeek
tags: 好观点 · → daily
📷 原图
展开 2 条中性
- <a id="atom-b3f4bce38bb2ae48"></a>2026-06-12
X·@aimlapi · Kimi K2.7-Code 发布 · 模型发布/产品发布
Kimi K2.7-Code is now available on AI/ML API!
tags: 好信源·好思考 · → daily
- <a id="atom-2e644c1a11491267"></a>2026-06-12
X·@VentureBeat · skipped DeepSWE · 技术路线
Moonshot skipped DeepSWE
tags: 好信源 · → daily
2. 算力资源 受限 vs 融资弹药充足
*topic: 数据中心算力/融资/供给产能 · 1 bull vs 1 bear*
🟢 bullish 侧:
- <a id="atom-1b4ed1302e26901f"></a>2026-06-14
X·@zephyr_z9 · access to GW class compute · 数据中心算力
DeepSeek, Moonshot will have access to GW class compute thanks to their war chests
tags: 好思考 · → daily
🔴 bearish 侧:
- <a id="atom-9d7f24ea23c19cbe"></a>2026-06-18
X·@teortaxesTex · relatively compute-poor · 供给产能
tags: 好观点 · → daily
展开 3 条中性
- <a id="atom-635da154ecdc64c9"></a>2026-05-06
X·@yicaichina · 融资领投方 · 融资
led by Meituan Longzhu
tags: 好信源 · → daily
📷 原图
- <a id="atom-cc01d0e06943e01c"></a>2026-05-19
X·@Techmeme · 计划进行香港IPO · 融资/IPO
Sources: Chinese AI startup Moonshot told investors it would revamp its corporate structure to pave the way for a Hong Kong IPO and comply with Beijing's rules
tags: 好信源 · → daily
value: date=未来时间 · direction=准备
🔗 因果传导 (causal map)
*Moonshot 在产业链上的传导关系. 边是群里/卖方陈述的因果 (非 AI 推断), 数字 = 几条 atom 支撑. 点 atom 溯源.*
↑ 上游·近期陈述 (1)
- 🟢 DeepSeek · 替代 · 1 次陈述
When DeepSeek figured out MLA that saves KV Cache by 93%, many major labs like ZAI and MoonShot adopted it.
🟦 客观事实 (facts) (22)
- <a id="atom-5d5708c1c57a4733"></a>🟦 2026-05-06
X·@yicaichina ⭐ · 融资规模 · 融资
Moonshot, the Chinese startup behind AI tool Kimi, is close to completing a USD2 billion fundraiser led by Meituan Longzhu
tags: 好数字·好信源 · → daily
value: qty=USD2 billion · date=2026-05
📷 原图
- <a id="atom-de6fc3622f314699"></a>🟦 2026-06-25
X·@EpochAIResearch · B2B销售岗位占比 · 营收
Moonshot B2B销售占比: 17%
tags: 好数字·好信源 · → daily
value: qty=17% · date=2026-06-25
[图: 展示MiniMax、Moonshot和Z.ai三家AI公司在销售与GTM招聘中B2B销售与营销岗位占比的对比条形图 — Z.ai B2B销售占比: 83%; MiniMax B2B销售占比: 24%; Moonshot B2B销售占比: 17%]
📷 原图
- <a id="atom-b3f4bce38bb2ae48"></a>🟦 2026-06-12
X·@aimlapi · Kimi K2.7-Code 发布 · 模型发布/产品发布
Kimi K2.7-Code is now available on AI/ML API!
tags: 好信源·好思考 · → daily
- <a id="atom-be66d3b19932302a"></a>🟦 2026-06-07
X·@Techmeme · 估值 · 融资/估值
Sources: Beijing-based Moonshot... is in talks to raise $1B+ that would value it at $30B, up from a $20B valuation
tags: 好数字 · → daily
value: qty=$30B · date=本轮融资
- <a id="atom-3c35662c73e35ccb"></a>🟦 2026-05-06
X·@yicaichina · 投后估值 · 融资
at a valuation of over USD20 billion
tags: 好数字 · → daily
value: qty=USD20 billion · date=2026-05
📷 原图
- <a id="atom-ac0f317f4b2e6992"></a>🟦 2026-06-10
X·@Techmeme · 重新考虑红筹结构 · 融资/监管政策
Chinese companies like Moonshot are reconsidering "red-chip" structures
tags: 好信源 · → daily
value: date=2026-06-10
- <a id="atom-cc01d0e06943e01c"></a>🟦 2026-05-19
X·@Techmeme · 计划进行香港IPO · 融资/IPO
Sources: Chinese AI startup Moonshot told investors it would revamp its corporate structure to pave the way for a Hong Kong IPO and comply with Beijing's rules
tags: 好信源 · → daily
value: date=未来时间 · direction=准备
- <a id="atom-635da154ecdc64c9"></a>🟦 2026-05-06
X·@yicaichina · 融资领投方 · 融资
led by Meituan Longzhu
tags: 好信源 · → daily
📷 原图
- <a id="atom-b956d8c9f8f53e8c"></a>🟦 2026-05-06
X·@yicaichina · 融资参与方 · 融资
with China Mobile and Citic-backed CPE also participating
tags: 好信源 · → daily
📷 原图
- <a id="atom-b01c0efbacfcbc1f"></a>🟦 2026-06-21
X·@jayair · 市场份额 · 市场份额
Moonshot 份额: 3.4%
tags: 好数字 · → daily
value: qty=3.4% · date=2026-06-21
[图: 开源模型作者代币市场份额对比图表,显示DeepSeek等模型的市场份额变化趋势及当前占比 — DeepSeek 份额: 65.0%; MiniMax 份额: 14.3%; Zhipu 份额: 8.1%; Xiaomi 份额: 7.2%; Moonshot 份额: 3.4%]
📷 原图
- <a id="atom-2e644c1a11491267"></a>🟦 2026-06-12
X·@VentureBeat · skipped DeepSWE · 技术路线
Moonshot skipped DeepSWE
tags: 好信源 · → daily
- <a id="atom-da24855dbfa0610b"></a>🟦 2026-06-05
X·@deedydas · 估值 · 估值
Moonshot - $20B
tags: 好数字 · → daily
value: qty=$20B
- <a id="atom-0f54bbd874a96cf6"></a>🟦 2026-05-10
X·@poezhao0605 · 融资额 · 融资
Moonshot closed $2 billion
tags: 好数字 · → daily
value: qty=$2 billion
- <a id="atom-f7537b340aadd968"></a>🟦 2026-04-24
X·@poezhao0605 · valuation · 估值
Moonshot at $18bn.
tags: 好数字 · → daily
value: qty=$18bn
- <a id="atom-227893003be78a15"></a>🟦 2026-04-19
X·@bookwormengr · 模型参数规模 · 模型发布/技术路线
Moonshot has 1T parameter Linear Attention model
tags: 好数字 · → daily
value: qty=1T · date=2026-04-19
📷 原图
- <a id="atom-8aa471b80ecf7159"></a>🟦 2026-06-23
X·@teortaxesTex · 创始人师从 · 人事关系
The founder of Moonshot is a student of the co-founder of Zhipu
tags: 好信源 · → daily
📷 原图
- <a id="atom-4fe34821115584df"></a>🟦 2026-06-22
X·@lucy_NFT · 后期尝试提取重建Claude推理轨迹 · 技术路线
In a later phase, Moonshot used a more targeted approach, attempting to extract and reconstruct Claude’s reasoning traces
tags: 好信源 · → daily
- <a id="atom-066d228d93332146"></a>🟦 2026-06-22
X·@bookwormengr · 被ANT识别为蒸馏攻击访问者 · 技术路线
ANT identified access by DeepSeek, MiniMax and Moonshot
tags: 好信源 · → daily
- <a id="atom-81da3352b759016e"></a>🟦 2026-06-21
X·@bookwormengr · 从旧开放模型过滤推理轨迹 · 技术路线
filtering reasoning trajectories from older open models also works too (e.g. GLM doing it from Moonshot models)
tags: 好信源 · → daily
- <a id="atom-5d72b1f9ed25e430"></a>🟦 2026-04-21
X·@teortaxesTex · 模型更新动态 · 模型发布
We've seen Moonshot release strong updates
tags: 好信源 · → daily
- <a id="atom-f6dc82dc9b2180b9"></a>🟦 2026-04-07
X·@bioshok3 · 被Anthropic识别为蒸馏行为方 · 竞争格局/技术路线
今年2月にはDeepSeek・Moonshot・MiniMaxの3社を蒸留行為で特定
tags: 好信源 · → daily
value: date=2025年2月
- <a id="atom-ea3bf6b7e661913b"></a>🟦 2026-04-22
X·@poezhao0605 · 押注于智能体编排 · 产品发布
Moonshot bet on agent orchestration.
tags: 好观点 · → daily
🟥 多头 takes (bullish) (16)
- <a id="atom-3545d463c3f04eab"></a>🟥 2026-04-26
X·@teortaxesTex [老化中] · K2.6 RL compute advantage over DeepSeek V4 · 技术路线
it's maybe up to 1 OOM above V4 in RL and like +10T tokens.
tags: 好数字·好观点 · → daily
value: qty=up to 1 OOM · direction=above
- <a id="atom-37a406f7424cb5fd"></a>🟥 2026-06-30
X·@tphuang · GLM-5.2 在 ZCode 上体验优于 DeepSeek V4 和 Kimi Code K2.7 · 模型发布/竞争格局
if you try GLM-5.2 on ZCode, it's an entirely different experience than V4 or even K2.7 with Kimi Code (different in a good way)
tags: 好观点·好思考 · → daily
- <a id="atom-346c43a4cf827737"></a>🟥 2026-04-29
X·@teortaxesTex · 表现 · 竞争格局
Moonshot is doing well
tags: 好信源 · → daily
📷 原图
- <a id="atom-1b4ed1302e26901f"></a>🟥 2026-06-14
X·@zephyr_z9 · access to GW class compute · 数据中心算力
DeepSeek, Moonshot will have access to GW class compute thanks to their war chests
tags: 好思考 · → daily
- <a id="atom-0fe1ab65d3a63d44"></a>🟥 2026-05-28
X·@teortaxesTex [老旧] · 推理智能体开发进度对比DeepSeek · 技术路线/竞争格局
They're consistently 4-6 months ahead of DeepSeek on the agentic stuff.
tags: 好观点 · → daily
value: qty=4-6个月 · direction=ahead
- <a id="atom-d71159c7e8f9f549"></a>🟥 2026-06-30
X·@tphuang · 即将发布的七月版本比 DeepSeek 完整版更令人期待 · 产品发布/仓位情绪
I'm far more looking forward to the next Moonshot release than the full version of V4
tags: 好观点 · → daily
- <a id="atom-48836216b5dfa2b5"></a>🟥 2026-06-25
X·@Presidentlin · shipping almost SOTA models · 技术路线/产品发布
Moonshot still shipping almost SOTA models
tags: 好观点 · → daily
- <a id="atom-0be8d69fcaf7808b"></a>🟥 2026-06-18
X·@teortaxesTex · best cultural fit · 合作客户
They have the best cultural fit
tags: 好观点 · → daily
- <a id="atom-a0ab793d53f69e72"></a>🟥 2026-06-16
X·@SanthProject · Q4 2026 表现预期 · 竞争格局
My prediction that q4 2026 moonshot will dominate is on track
tags: 好观点 · → daily
value: direction=dominate · date=Q4 2026
- <a id="atom-4a4ffcb9b0d82cb5"></a>🟥 2026-06-16
X·@Presidentlin · 近期表现评价 · 竞争格局
the Moonshot team have been doing a very good job
tags: 好观点 · → daily
- <a id="atom-352e5115fe8d120d"></a>🟥 2026-06-14
X·@_xjdr · k2.7 表现 · 模型性能_
k2.7 has been extremely impressive so far
tags: 好观点 · → daily
- <a id="atom-a99414af21ea504d"></a>🟥 2026-06-14
X·@_xjdr · k2.6 表现 · 模型性能_
as was k2.6 before it
tags: 好观点 · → daily
- <a id="atom-08efb11da8528f5a"></a>🟥 2026-06-14
X·@akhileshutup · k2.7 与 k2.6 对比 · 模型性能
it's far less verbose wrt k2.6, which i felt had overthinking issues
tags: 好观点 · → daily
- <a id="atom-3e2eae76e8a583b0"></a>🟥 2026-06-03
X·@TheAhmadOsman [老旧] · 中国实验室 tier 评级 · 竞争格局
S-Tier Chinese Labs: Moonshot and DeepSeek
tags: 好观点 · → daily
value: qty=S-Tier · date=2026-06-03
- <a id="atom-04c47284713c121e"></a>🟥 2026-05-01
X·@teortaxesTex [老化中] · 通过强化学习可在推理方面超越DeepSeek · 竞争格局/技术路线
for now Moonshot or w/e can outpace them with muh RL
tags: 好观点 · → daily
- <a id="atom-29647764f0a232e1"></a>🟥 2026-04-26
X·@teortaxesTex [老化中] · K2.6 model strength relative to DeepSeek V4 · 技术路线
I think K2.6 is stronger than V4 though.
tags: 好观点 · → daily
value: direction=stronger
🟥 空头 takes (bearish) (7)
- <a id="atom-625dbb6b0f15aa11"></a>🟥 2026-04-18
X·@zephyr_z9 [老化中] · model open source justification · 竞争格局
Very hard to justify model open sourcing as a small public company
tags: 好思考 · → daily
- <a id="atom-62399a513b510678"></a>🟥 2026-04-18
X·@zephyr_z9 · M3 model open source status · 技术路线
M3 won't get open sourced
tags: 好观点 · → daily
- <a id="atom-9d7f24ea23c19cbe"></a>🟥 2026-06-18
X·@teortaxesTex · relatively compute-poor · 供给产能
tags: 好观点 · → daily
- <a id="atom-4ba3e96389458781"></a>🟥 2026-05-24
X·@thegenioo · Kimi K2.6 表现评价 · 产品发布
Honestly I think Moonshot fumbled big time with Kimi K2.6
tags: 好观点 · → daily
- <a id="atom-ad2916bda34df3e9"></a>🟥 2026-05-24
X·@thegenioo · Kimi K2.6 问题 — 过度思考循环 · 技术路线
It just overthinks the hell out of things and gets stuck in long, endless thinking loops
tags: 好观点 · → daily
- <a id="atom-6e60dd949bea7a70"></a>🟥 2026-05-24
X·@thegenioo · Kimi K2.6 问题 — 速度慢 · 技术路线
It is unbelievably slow, like really slow
tags: 好观点 · → daily
- <a id="atom-96cda960c60d80a9"></a>🟥 2026-04-28
X·@teortaxesTex [老化中] · 战略窗口期 · 竞争格局
there is a small window now for Minimax/Moonshot/Zhipu to make excellent moves and secure their stake in the future, otherwise they're getting squeezed to death between Alibaba&ByteDance and DeepSeek
tags: 好观点 · → daily
📷 原图
🟥 中性 takes (neutral) (1)
- <a id="atom-7ba35d9c51439368"></a>🟥 2026-05-19
X·@Sino_Market · 考虑重组以为IPO铺路 · IPO
MOONSHOT CONSIDERING RESTRUCTURING TO PAVE WAY FOR IPO, SOURCES SAY
tags: 好信源 · → daily
⏱ 时间轴 (近 20)
- 🟥 2026-06-30 ·
narrative · GLM-5.2 在 ZCode 上体验优于 DeepSeek V4 和 Kimi Code K2.7 · → daily
- 🟥 2026-06-30 ·
narrative · 即将发布的七月版本比 DeepSeek 完整版更令人期待 · → daily
- 🟥 2026-06-25 ·
forecast · shipping almost SOTA models · → daily
- 🟦 2026-06-25 ·
fact · B2B销售岗位占比 · → daily
- 🟦 2026-06-23 ·
fact · 创始人师从 · → daily
- 🟦 2026-06-22 ·
fact · 后期尝试提取重建Claude推理轨迹 · → daily
- 🟦 2026-06-22 ·
fact · 被ANT识别为蒸馏攻击访问者 · → daily
- 🟦 2026-06-21 ·
fact · 市场份额 · → daily
- 🟦 2026-06-21 ·
fact · 从旧开放模型过滤推理轨迹 · → daily
- 🟥 2026-06-18 ·
narrative · best cultural fit · → daily
- 🟥 2026-06-18 ·
narrative · relatively compute-poor · → daily
- 🟥 2026-06-16 ·
forecast · Q4 2026 表现预期 · → daily
- 🟥 2026-06-16 ·
narrative · 近期表现评价 · → daily
- 🟥 2026-06-14 ·
forecast · access to GW class compute · → daily
- 🟥 2026-06-14 ·
narrative · k2.7 表现 · → daily
- 🟥 2026-06-14 ·
narrative · k2.6 表现 · → daily
- 🟥 2026-06-14 ·
narrative · k2.7 与 k2.6 对比 · → daily
- 🟦 2026-06-12 ·
fact · Kimi K2.7-Code 发布 · → daily
- 🟦 2026-06-12 ·
fact · skipped DeepSWE · → daily
- 🟦 2026-06-10 ·
fact · 重新考虑红筹结构 · → daily
*以下是 Chinese AI labs 板块的信号 (Moonshot 是成员), 非直接针对 Moonshot — 供板块上下文参考*
展开板块信号
- 🟥▲ 2026-06-13 [twitter] · 价格竞争力
There’s also really competitive offering from chinese labs out, quality is usually a year behind latest gen but for 10%
- 🟦▲ 2026-06-07 [twitter] · gap 10 years ago
Chinese labs were 10 years behind (per The Economist), now they are 6 months
- 🟥▼ 2026-06-10 [twitter] · 面临美国网络攻击风险
They want to rugpull Chinese labs – no, the entire Chinese IT, if not the whole infrastructure of the society – near the
- 🟥▲ 2026-06-07 [twitter] · gap to frontier closed models
the difference is at max 6 months
- 🟥▲ 2026-06-07 [twitter] · fundamental capability gap
There is no fundamental human resource capability gap. It is compute and $$$ and datasets.
- 🟥▼ 2026-06-04 [twitter] · 停止发布开放权重模型的影响
If they stop, the frontier falls further and further behind to those who want to use local/fine-tuned models
- 🟥▲ 2026-05-31 [twitter] · resource advantages relative to US labs
cheaper energy, more abundant AI researchers and the ability to distill and learn from leading US models, but less capab
- 🟥▼ 2026-07-01 [twitter] · would not economically survive Anthropic competition
Chinese AI labs would not economically survive the juggernaut of Anthropic - unless China took drastic steps
← 实体目录 · 系统日志