[AINews] Claude Haiku 5.5:同价位表现优于 GPT-6 Luna
英文标题:[AINews] Claude Haiku 5.5 — better than GPT-6 Luna at the same pricing
Anthropic 发布 Claude Haiku 5.5,这是约一年来首个 Haiku 档位更新,定价与 OpenAI 的 GPT-6 Luna 相同,并同步下调 Sonnet 5.5 缓存读取价格、向订阅用户提供 API 额度。Artificial Analysis 评测显示,Haiku 5.5 在最高 effort 下的 Intelligence Index 为 43,略高于 GPT-6 Luna 的 38,但其输出 token 用量约为 Luna 的 3 倍,实际成本优势仍需结合任务和上下文长度判断。该模型支持 effort settings、adaptive thinking 和 100 万 token 上下文,被定位为 Opus 5.5 或 Sonnet 5.5 驱动的 Agent 工作流中的低成本高吞吐子 Agent。Cursor、GitHub Copilot、Devin 等合作伙伴已在发布当天接入,显示小模型正成为多模型 Agent 系统中承担摘要、压缩、数据库查询和执行类任务的重要一层。
1.5 KB
2.2 KB
1.1 KB
176.3 KB
96.5 KB
292.2 KB
中文
距 Anthropic 发布 Haiku 4.5 已过去约一年。随着 Sonnet、Opus 和 Fable 相继迭代至 5.5,Haiku 看起来似乎有些被遗忘,尤其是在 OpenAI 与 Astra、Sol 6 一同发布 Luna 6 之后。
不过,它终于来了,而且这次更新令人欢迎。更多内容见下方摘要。
2026/10/06—2026/10/07 的 AI 新闻。我们检查了 12 个 subreddit、544 个 Twitter 账号,未再检查 Discord。通过 AINews 网站可搜索往期全部内容。提醒一下:AINews 现已成为 Latent Space 的一个栏目。你可以选择订阅或取消订阅不同的邮件发送频率!
头条:Anthropic 发布 Claude Haiku 5.5
Anthropic 发布 Claude Haiku 5.5,这是约一年来首次更新 Haiku 档位模型。其定价与 OpenAI GPT-6 Luna 持平;Anthropic 同日还下调了 Sonnet 5.5 和订阅计划的价格。
-
发布前信号。 @scaling01 在公告前发布“祝 Haiku 5.5 日快乐”。@kimmonismus 表示,该模型已出现在 Claude Code 更新中,并预测会采用“Luna 定价”。随后,他在官方发布前提前公布价格(@kimmonismus),并在模型上线时予以确认(@kimmonismus)。
-
官方发布。 @claudeai 和 @AnthropicAI 将其称为“我们发布过最便宜、最快、能力最强的小型模型”,平均运行成本约比 Haiku 4.5 低 75%。
-
可用性与目标用途。 该模型已在 Claude Platform 与 Claude Code 中上线(@ClaudeDevs)。Anthropic 将其定位为与 Opus 5.5 或 Sonnet 5.5 配套使用的子 Agent,适用于摘要、上下文压缩和数据库查询等高吞吐、成本敏感型工作。@mikeyk 将这种分工描述为:“Opus 负责重度思考,Haiku 负责高吞吐工作。”
-
分级定价(@ClaudeDevs):
| 提示词长度 | 每 100 万 token 输入 / 输出价格 | 每 100 万 token 缓存读取价格 |
|---|---|---|
| 低于 100K token | $0.10 / $0.50 | $0.01 |
| 高于 100K token | $0.50 / $2.50 | $0.05 |
-
Sonnet 5.5 降价。 缓存读取价格从每 100 万 token $0.20 降至 $0.10。Anthropic 表示,这使 Sonnet 5.5 在多数长时间运行或 Agent 类任务中的成本降低约 20%(@claudeai)。
-
订阅用户 API 额度。 Max 5x($100)、Max 20x($200)和 Team(最高 $500,可共享)套餐现在均附带每月 Claude Platform API 额度。这些额度适用于任何模型,包括 Haiku 5.5,也可用于第三方 harness(@ClaudeDevs)。
-
同日 SDK 更新。 计算机使用和浏览器使用工具集现已内置于 Python 和 TypeScript Claude SDK。SDK 可运行动作循环,并将点击和按键操作发送至 browser_use、Browserbase、E2B 或 Daytona 的驱动程序,开发者无需再自行编写该循环(@ClaudeDevs,快速开始)。
-
首日合作伙伴接入:
-
Cursor:@cursor_ai 称,在较短请求中,该模型价格“比 Haiku 4.5 低 10 倍”,并发布了 CursorBench 对比结果(@cursor_ai)。
-
VS Code 中的 GitHub Copilot:@code 报告称,它“在许多编程任务上匹敌 Claude Sonnet 5,同时使用更少 token 和步骤”。
-
Devin:@cognition 报告,该模型在 FrontierCode 1.1 上获得 58.4%,领先 Sonnet 5,而每项任务成本约为后者的八分之一;并建议在 Fusion 中将其作为 Opus 5.5 主 Agent 之下的“搭档”。
-
Arena:已加入 Agent Arena、Code Arena WebDev、Text、Document 和 Vision,评分尚待公布(@arena)。
-
OpenDocRouter:同日接入,详见下文。
-
@ArtificialAnlys 发布了最详细的第三方数据(按评测拆分、对比页面)。
-
Intelligence Index:43 分,在最高 effort 下较上一代 Haiku 提高 26 分。
-
略高于 GLM-5.3 Flash(42)、Gemini 3.8 Flash(41)和 GPT-6 Luna(38)。
-
与 Kimi K3(44)相近;后者是一个拥有 2.8T 参数的开放权重模型。
-
比最高 effort 下的 Claude Sonnet 5.5(56)低 13 分。
-
-
新增控制项。 这是首个支持 Anthropic effort settings 和 adaptive thinking 的 Haiku。
-
Token 用量是主要注意事项。
-
在最高 effort 下,它每项 Intelligence Index 任务约使用 16.2 万输出 token,约为最高 effort 下 GPT-6 Luna(约 5 万)的 3 倍。
-
从
xhigh升至max,约多消耗 1.8 倍 token,仅增加 2 分。 -
在相同得分下,它仍更为冗长:高 effort 的 Haiku 5.5 使用约 5.5 万 token 得分 38,而最高 effort 的 Luna 使用约 5 万 token 也得分 38。更低 effort 设置下,差距会进一步扩大。
-
-
成本数据仍为暂定值。 Artificial Analysis 尚未将超过 100K token 后的 5 倍价格阶梯纳入模型。后续将提供单任务成本数据。
-
AA-Briefcase(私有 Agent 知识工作评测):1578 Elo。该成绩超过 Kimi K3 和 GLM-5.3,与最高设置下的 Muse Spark 1.3 相当。
-
Terminal-Bench 4.0:33%,高于 Haiku 4.5 的 0%。
-
知识准确性与幻觉(AA-Omniscience):
| 模型 | 准确率 | 幻觉率 |
|---|---|---|
| Haiku 5.5 | 36% | 40% |
| Gemini 3.8 Flash | 55% | 55% |
| GPT-6 Luna | 44% | 77% |
Haiku 较低准确率的一部分原因,是它更愿意承认自己不知道答案。
-
AutomationBench-AA:35%,相比之下 Luna、Gemini 3.8 Flash 和 GLM-5.3 Flash 的成绩为 53%—60%。一个发布前安全性 bug 导致模型过度拒答。Anthropic 正在修复该问题;Artificial Analysis 将重新运行评测,并预计分数会上升。
-
规格:
-
100 万 token 上下文,较 Haiku 4.5 的 20 万 token 提升。
-
支持文本与图像输入、文本输出。
-
5 分钟缓存写入价格为每 100 万 token $0.125;超过 100K token 时为 $0.625。
-
-
@ShayneRedford 总结了 Anthropic 报告的提升:
-
OSWorld(计算机使用):15% → 72%。
-
TerminalBench:0% → 39%。这与 Artificial Analysis 在 Terminal-Bench 4.0 上独立测得的 33% 不同。
-
知识工作和推理能力提升 10%—50%。
-
在其中多数项目上超过 Luna。
-
100 万 token 上下文,最大输出约 1.2 万 token。
-
-
@alexalbert__(Anthropic)强调,Haiku 4.5 于 2025 年 10 月 15 日发布,因此比较跨度不足一年。
-
@TheRundownAI 报告称,它在“多种基准测试”中击败 GPT-6 Luna。
-
文档解析(独立 ParseBench 测试):
-
@LoganMarkewich:整体接近 Luna,表格处理略好,但图表理解较弱。
-
@jerryjliu0:每 1,000 页约 $1.2。以其价格而言,表格和阅读顺序处理良好;图表、语义格式及边界框能力较弱。
-
-
轶事反馈:
-
@simonw 撰写了定价说明,并运行了自己的“骑自行车的鹈鹕”测试。他表示,该模型比 Haiku 4.5“好太多了”,而 Haiku 4.5 的价格高 10 倍(对比)。
-
@AI_Screening 表示,Haiku 5.5 在 Three.js 斑马模拟上“完胜” Luna,但这只是单条提示词测试,并非系统性评估。
-
看多观点
-
@kimmonismus:“比 GPT-6-Luna 强得多,更接近 Sonnet 5.5……便宜又聪明。”
-
@theo 喜欢分级定价:对于低于 100K token 的请求仅收取五分之一价格,“非常明确地说明了这个模型的用途”。他还认为,看到 Anthropic 模型在成本/智能图表上“如此靠左”很引人注目(@theo),并在直播中讨论了此次发布(@theo)。
-
@draecomino:“最高 effort 下的 Haiku 表现得像一个前沿模型。”
-
@kipperrii 认为,小型廉价模型如今更为重要,因为它们已经能够完成“大量有用的工作”。
-
@scaling01 称其“便宜得离谱”,之后又表示(@scaling01):“5.5 系列模型看起来很不错。”
-
来自 Anthropic 阵营的 @NotTomBrown(“小而强大”)和 @edwinarbus 都发布了庆祝性评论。
竞争格局解读
-
此次发布被广泛解读为瞄准 OpenAI 的 GPT-6 Luna:“GPT-6 Luna 安息吧”(@dejavucoder),“该狠狠干 Luna 了”(@scaling01)。
-
@kimmonismus 将 Sonnet 缓存读取降价解读为 Anthropic 正在向 OpenAI 施压。谈及 API 额度时,他补充说:“OpenAI,该你了。”(@kimmonismus)
-
@teortaxesTex 表示,Anthropic 目前拥有“所有实验室中最深厚的产品线”:Haiku/Sonnet/Opus/Fable 加上 Mythos;相比之下 OpenAI 为 Luna/Sol/Astra。不过,他仍认为 Anthropic“较不重视产品”,并将此解读为该公司在 2026 年拥有大量余裕的迹象。
-
ThursdAI 的 @altryne 质疑中间档位的必要性:“当 Opus 价格昂贵时,Sonnet 是合理的。”该节目计划讨论 Haiku 5.5(@thursdai_pod)。
注意事项(大多来自数据,而非强烈批评)
-
标价可能夸大实际节省。
-
较高 token 用量(最高 effort 下约为 Luna 的 3 倍)会抵消部分按 token 计算的折扣。
-
超过 100K token 的提示词需支付 5 倍价格,这对长上下文 Agent 循环很重要。
-
这两项因素很可能解释了 Cursor 所称“短请求便宜 10 倍”与 Anthropic 所称“平均便宜 75%”之间的差异。
-
-
事实性记忆弱于 Gemini 3.8 Flash 和 Luna。
-
过度拒答 bug目前压低了自动化分数。
-
**尚未完成评测:**Arena 分数仍待公布,独立的单任务成本数据则在等待支持分级定价。
-
自 2025 年 10 月以来,Haiku 4.5 一直是 Anthropic 的小模型。在此期间,OpenAI 的 GPT-6 Luna、Gemini 3.8 Flash、GLM-5.3 Flash 和 DeepSeek V4.1 Flash 一直在低成本市场竞争。
-
Haiku 5.5 的标价与 Luna 完全相同,并新增 effort 控制与 100 万 token 上下文。
-
它被明确定位为多模型 Agent harness 中的低价执行者:Claude Code 子 Agent、Devin Fusion、Copilot 子 Agent,以及上下文压缩或摘要步骤。
-
Anthropic 将其与 Sonnet 缓存读取降价、订阅 API 额度捆绑发布,构成了一次围绕 Agent 经济性的协同推进。这一动作落在行业对于 token 账单的更广泛讨论之中(见下方 @theo)。
OpenAI 的 722 份数学手稿:规模、效率与余波
English / 原文
It’s been about a year since Anthropic shipped Haiku 4.5, and with successive launches of Sonnet and Opus and Fable up to 5.5 it was seeming a little forgotten, especially as OpenAI launched Luna 6 alongside Astra and Sol 6.
Well, it’s here, and it’s a welcome update. More in the summary below.
[Artificial Analysis@ArtificialAnlysAnthropic has released Claude Haiku 5.5, scoring 43 on the Artificial Analysis Intelligence Index - up 26 points one year after the last Haiku release Haiku 5.5 is the first Haiku model with Anthropic’s effort settings and adaptive thinking, and Anthropic has introduced tiered …7:12 PM · Oct 7, 2026 · 152K Views
92 Replies · 111 Reposts · 1.67K Likes](https://x.com/ArtificialAnlys/status/2107911905822351609)
AI News for 10/06/2026-10/7/2026. We checked 12 subreddits, 544 Twitters and no further Discords. AINews’ website lets you search all past issues. As a reminder, AINews is now a section of Latent Space. You can opt in/out of email frequencies!
Top Story: Anthropic releases Claude Haiku 5.5
Anthropic shipped Claude Haiku 5.5, its first Haiku-tier update in about a year. It is priced to match OpenAI’s GPT-6 Luna, and Anthropic cut prices on Sonnet 5.5 and its subscription plans the same day.
-
Pre-launch signals. @scaling01 posted “happy Haiku 5.5 day” before the announcement. @kimmonismus said the model had already appeared in a Claude Code update and predicted “Luna-pricing.” He then posted the pricing ahead of the official post (@kimmonismus) and confirmed when it went live (@kimmonismus).
-
Official launch. @claudeai and @AnthropicAI called it “the cheapest, fastest, and most capable small model we’ve ever released,” costing about 75% less to run than Haiku 4.5 on average.
-
Availability and intended use. It is live on the Claude Platform and in Claude Code (@ClaudeDevs). Anthropic positions it as a subagent paired with Opus 5.5 or Sonnet 5.5, for high-volume, cost-sensitive work such as summaries, compactions and database queries. @mikeyk described the split as “Opus does the heavy thinking, Haiku does the high-volume work.”
-
Tiered pricing (@ClaudeDevs):
Prompt lengthInput / output per 1M tokensCache reads per 1MUnder 100K tokens$0.10 / $0.50$0.01Over 100K tokens$0.50 / $2.50$0.05
-
Sonnet 5.5 price cut. Cache reads were halved from $0.20 to $0.10 per 1M tokens. Anthropic says this makes Sonnet 5.5 about 20% cheaper on most long-running or agentic work (@claudeai).
-
API credits for subscribers. Monthly Claude Platform API credits now come with Max 5x ($100), Max 20x ($200) and Team (up to $500, pooled). They work on any model, including Haiku 5.5, and in third-party harnesses (@ClaudeDevs).
-
Same-day SDK update. Computer-use and browser-use toolsets are now built into the Python and TypeScript Claude SDKs. The SDK runs the action loop and sends clicks and keystrokes to drivers from browser_use, Browserbase, E2B or Daytona, so developers no longer write that loop themselves (@ClaudeDevs, quickstart).
-
Partner rollouts on day one:
-
Cursor: @cursor_ai claims “10x less than Haiku 4.5” on shorter requests and published CursorBench comparisons (@cursor_ai).
-
GitHub Copilot in VS Code: @code reports it “matched Claude Sonnet 5 on many coding tasks while using fewer tokens and steps.”
-
Devin: @cognition reports 58.4% on FrontierCode 1.1, ahead of Sonnet 5 at roughly one-eighth the cost per task, and recommends it as a “sidekick” under an Opus 5.5 lead in Fusion.
-
Arena: added to Agent Arena, Code Arena WebDev, Text, Document and Vision, with scores pending (@arena).
-
OpenDocRouter: added the same day (details below).
-
@ArtificialAnlys published the most detailed third-party numbers (per-eval breakdown, comparison page).
-
Intelligence Index: 43 at max effort, up 26 points from the previous Haiku.
-
Slightly ahead of GLM-5.3 Flash (42), Gemini 3.8 Flash (41) and GPT-6 Luna (38).
-
Comparable to Kimi K3 (44), a 2.8T-parameter open-weights model.
-
Trails Claude Sonnet 5.5 at max effort (56) by 13 points.
-
-
New controls. This is the first Haiku with Anthropic’s effort settings and adaptive thinking.
-
Token usage is the main caveat.
-
At max effort it uses about 162k output tokens per Index task, roughly 3x GPT-6 Luna at max (~50k).
-
Going from xhigh to max adds 2 points for about 1.8x the tokens.
-
At equal score it is still more verbose: Haiku 5.5 at high effort scores 38 using ~55k tokens, versus Luna at max scoring 38 with ~50k. The gap widens at lower effort settings.
-
-
Cost figures are provisional. Artificial Analysis does not yet model the 5x price step above 100K tokens. Cost-per-task numbers will follow.
-
AA-Briefcase (private agentic knowledge-work eval): 1578 Elo. That is ahead of Kimi K3 and GLM-5.3, and comparable to Muse Spark 1.3 at max.
-
Terminal-Bench 4.0: 33% , up from 0% for Haiku 4.5.
-
Knowledge versus hallucination (AA-Omniscience):
ModelAccuracyHallucination rateHaiku 5.536%40%Gemini 3.8 Flash55%55%GPT-6 Luna44%77%
Part of Haiku’s lower accuracy comes from being more willing to say it doesn’t know.
-
AutomationBench-AA: 35% , versus 53–60% for Luna, Gemini 3.8 Flash and GLM-5.3 Flash. A pre-release safety bug caused the model to over-refuse. Anthropic is working on a fix, and Artificial Analysis will re-run the eval and expects the score to rise.
-
Specs:
-
1M-token context, up from 200k for Haiku 4.5.
-
Text and image input, text output.
-
5-minute cache writes cost $0.125 per 1M tokens ($0.625 above 100K).
-
-
@ShayneRedford summarized Anthropic’s reported jumps:
-
OSWorld (computer use): 15% → 72%.
-
TerminalBench: 0% → 39%. This differs from Artificial Analysis’s independent 33% on Terminal-Bench 4.0.
-
10–50% gains in knowledge work and reasoning.
-
Beats Luna on most of these.
-
1M context with roughly 12k max output tokens.
-
-
@alexalbert__ (Anthropic) stressed that Haiku 4.5 shipped Oct 15, 2025, so the comparison spans less than a year.
-
@TheRundownAI reported that it beats GPT-6 Luna “across a variety of benchmarks.”
-
Document parsing (independent, on ParseBench):
-
@LoganMarkewich: overall close to Luna, slightly better on tables, worse on chart understanding.
-
@jerryjliu0: about $1.2 per 1,000 pages. Good at tables and reading order for the price; weaker on charts, semantic formatting and bounding boxes.
-
-
Anecdotal:
-
@simonw wrote pricing notes and ran his pelican-on-a-bicycle test. He says it is “SO MUCH better” than Haiku 4.5, which costs 10x more (comparison).
-
@AI_Screening says Haiku 5.5 “cooked” Luna on a Three.js zebra simulation (single prompt, not systematic).
-
Bullish
-
@kimmonismus: “Way better than GPT-6-Luna, close[r] to Sonnet 5.5… Cheap and smart.”
-
@theo likes the tiered pricing: charging a fifth of the price under 100K tokens “makes it really clear what the model is for.” He also found it striking to see an Anthropic model “so far to the left on the cost/intelligence charts” (@theo) and covered the launch on stream (@theo).
-
@draecomino: “Haiku at max effort performs like a frontier model.”
-
@kipperrii argues small, cheap models matter more now that they can do “a ton of useful things.”
-
@scaling01 (”cheap af”) and later (@scaling01): “5.5 models are looking good.”
-
@NotTomBrown (”small but mighty”) and @edwinarbus, both from the Anthropic side, posted celebratory notes.
Competitive framing
-
The launch is widely read as aimed at OpenAI’s GPT-6 Luna: “rip gpt 6 luna” (@dejavucoder), “time to cook Luna” (@scaling01).
-
@kimmonismus framed the Sonnet cache-read cut as Anthropic pressuring OpenAI. On the API credits he added, “OpenAI: your turn” (@kimmonismus).
-
@teortaxesTex says Anthropic now has “the deepest product lineup of all labs” (Haiku/Sonnet/Opus/Fable plus Mythos) versus OpenAI’s Luna/Sol/Astra. He still thinks Anthropic “cares less about products,” which he reads as a sign of how much slack it has had through 2026.
-
ThursdAI’s @altryne questioned the middle tier: “Sonnet made sense when Opus was expensive.” The show plans to cover Haiku 5.5 (@thursdai_pod).
Caveats (mostly from the data, not loud critics)
-
The headline price may overstate real savings.
-
Heavy token use (about 3x Luna at max effort) offsets part of the per-token discount.
-
Prompts over 100K tokens pay 5x more, which matters for long-context agent loops.
-
These two effects likely explain why Cursor’s “10x cheaper on short requests” differs from Anthropic’s “75% cheaper on average.”
-
-
Factual recall is weaker than Gemini 3.8 Flash and Luna.
-
The over-refusal bug currently depresses the automation score.
-
Not yet benchmarked: Arena scores are pending, and independent cost-per-task figures await tiered-pricing support.
-
Haiku 4.5 has been Anthropic’s small model since October 2025. Since then, OpenAI’s GPT-6 Luna, Gemini 3.8 Flash, GLM-5.3 Flash and DeepSeek V4.1 Flash have competed at the low-cost end.
-
Haiku 5.5 matches Luna’s sticker price exactly. It adds effort control and a 1M context.
-
It is explicitly positioned as the cheap worker inside multi-model agent harnesses: Claude Code subagents, Devin Fusion, Copilot subagents, and compaction or summarization steps.
-
Anthropic paired it with Sonnet cache-read cuts and subscription API credits, a coordinated push on agent economics. That push lands amid wider debate over token bills (see @theo below).
OpenAI’s 722 Math Manuscripts: Scale, Efficiency and Fallout