Claude Sonnet 5.5

發布於 2026-09-28

chat程式碼推理工具呼叫圖像理解提示詞快取

Claude Sonnet 5.5 是 Anthropic 在速度與智慧之間取得最佳平衡的模型,於 2026 年 9 月 28 日發布,輸入與輸出價格與 Claude Sonnet 5 相同:每百萬輸入 token $2、每百萬輸出 token $10,提示詞快取讀取為輸入價的 5%(每百萬 $0.10,是 Sonnet 5 的一半),5 分鐘快取寫入 $2.50,1 小時快取寫入 $4。

輸入
文字 圖像 $2/M
輸出
文字 $10/M
快取讀取
$0.1/M
上下文
1M
相較 GPT-4o
便宜約 60%
知識截止
2026-06

Benchmark 成績

Claude Sonnet 5.5:公布了 10 項,但每一項都沒有夠多其他模型的成績,無法比較。

Claude Sonnet 5.5 其他有成績的模型 其他模型平均 ★ 沒有模型分數更高
SWE-sweep Hydrogen: Antigravity CLI 1.4.8. Peers: official leaderboard
3.1%
OSWorld 2.1 82 offline tasks, partial score. Hydrogen and Sonnet: best of 3. Terra: system card
83.9%
Agents’ Last Exam v1.1 Oct 17 task set. Hydrogen: Antigravity CLI. Sonnet: ALE-Claw. Terra: official leaderboard
60.4%
HLE-Diamond Artificial Analysis leaderboard, November report
53.3%
Vending-Bench 2 (USD) Andon Labs leaderboard, Hydrogen report · 9107-14549
11165
Chartography No tools. Hydrogen and Terra: Google evaluation. Sonnet: system card
61.6%

供應商公布: Google

價格在同類中的位置

價格在 79 個同類模型中的位置

輸入$2/M
$0.1 · Claude Haiku 5.5 GPT-5.4 Pro · $30
輸出$10/M
$0.275 · DeepSeek V4 Flash GPT-5.4 Pro · $180
快取讀取$0.1/M
$0.006 · DeepSeek V4.1 Flash Claude Fable 5 · $1

這條線標出這個模型的價格,在 Synthorai 的同類模型裡落在什麼位置;兩端分別是最便宜和最貴的模型。這裡列的是基礎價,批次、區域和快取寫入的折扣請見價格頁。

規格與限制

Token

上下文視窗(供應商規格) 1,000,000
最大輸出(供應商規格) 128,000
知識截止 2026-06

提示詞快取

快取方式 顯式(需主動開啟)
最小前綴 512 供應商預設 1,024
存活時間 預設 5 分鐘,可選 1 小時
寫入成本 1.25x (5m) / 2x (1h)

思考

供應商參數 thinking.type
可接受的值 adaptive (default) · between_tools
預設值 adaptive, effort high 請求未指定時生效
可關閉 不支援
思考行為 Adaptive thinking is on by default. The lowest setting, between_tools, turns off up-front thinking and works at high effort or below; thinking {"type": "disabled"} and a manual {"type": "enabled", "budget_tokens": N} both return a 400 error.
參數 reasoning_effort
參數值 minimal · low · medium · high 這是閘道的參數,實際以上方的供應商對應為準

模型

模態 文字 + 圖像 → 文字
  • Same input and output prices as Claude Sonnet 5
  • 1M context at standard pricing with no long-context tier
  • prompt-cache reads cost 0.05x input ($0.10/M) rather than the usual 0.1x
  • minimum cacheable prompt 512 tokens
  • setting temperature, top_p or top_k to a non-default value returns a 400
  • forced tool use returns an error
  • text between tool calls comes back in thinking blocks
  • up to 300k output tokens on the Message Batches API with the output-300k-2026-03-24 beta header

依據 模型規格來源 ↗

一段提示詞,經閘道實測

提示詞 Explain to a CFO, in exactly three sentences, why our LLM bill doubled after we enabled reasoning. 檢查 恰好 3 句

Claude Sonnet 5.5 通過 · 3 sentences

Enabling reasoning makes the model "think" before it answers, and that internal thinking is generated as extra text (reasoning tokens) that we are billed for at the same rate as output tokens, even though users never see it. Output tokens also cost several times more than input tokens, so a single request that used to produce a short answer can now generate many times more billable text behind the scenes. Because our request volume and prompts didn't change, the doubling comes almost entirely from this hidden thinking, and we can bring it down by limiting reasoning effort, turning it off for simple tasks, and reserving it for the workloads where the accuracy gain justifies the cost.

輸出 205 tok 延遲 3.6 s

是否照指令做(剛好三句,數得出來)、語氣是否貼合對象(寫給 CFO 看),以及下方 token 計量顯示出的隱藏思考計費落差。

提示詞 This function is supposed to return the longest run of consecutive equal items, but callers report it is off by one on single-element inputs and misses runs that end at the last index. Fix it and explain the bug in one paragraph. 檢查 修正後通過測試

Claude Sonnet 5.5 通過 · 8/8 cases

**The bug:** `cur` counts the number of *adjacent equal pairs* rather than the number of items in the run. A run of length k contains only k−1 such pairs, so every result is one too small. The counter also starts at 0 and the loop starts at index 1, so a single-element list never updates `best` and returns 0 instead of 1. (An empty list should return 0.) In addition, `cur` resets to 0 on a mismatch, when it should reset to 1, because the item at `xs[i]` begins a new run of length 1. The claim that it "misses runs ending at the last index" is really the same off-by-one: the final run is counted, but as one less than its true length. Because `best` is updated inside the loop on every extension, the last run isn't skipped. […]

輸出 444 tok 延遲 6.7 s

修正是否真的正確(跑得起來)、說明的資訊密度,以及在範圍明確的任務上的 token 效率。

提示詞 Extract a JSON object with fields {company, quarter, revenue_usd_m, growth_pct, guidance} from this text. Use null for anything not stated; add no other fields. 檢查 有效的 JSON,完全符合 schema

Claude Sonnet 5.5 通過 · 5/5 fields, guidance null

Here is the extracted JSON: ```json { "company": "Meridian Robotics", "quarter": "September quarter", "revenue_usd_m": 412, "growth_pct": 11, "guidance": null } ``` Notes on the choices: - **quarter**: The text says "September quarter" without specifying a fiscal year or Q-number, so I kept the wording as stated. - **revenue_usd_m**: The text gives "412" for the topline, with no currency or unit stated explicitly. I've taken it as 412 million USD per the field definition, but the text does not confirm that, and the figure is constant-currency. - **growth_pct**: "Eleven and change" means slightly above 11, so 11 is an approximation. The exact figure isn't given. - **guidance**: Guidance is withheld pending the Q4 close, so no guidance is stated, hence null.

輸出 300 tok 延遲 3.5 s

是否遵守 schema(不自創欄位)、面對幻覺誘因的表現(題目明確不給 guidance),以及結構化輸出路徑的差異。

提示詞 Write a 120-word product announcement for an API gateway feature that caches prompts across providers. Forbidden words: "seamless", "unlock", "game-changer", "revolutionize", "empower". Exactly one sentence must be a question. 檢查 120 詞,0 個禁用詞

Claude Sonnet 5.5 通過 · 122 words, 0 banned, 1 question

**Introducing Cross-Provider Prompt Caching** Your API gateway now caches prompts across every connected LLM provider, whether you route traffic to OpenAI, Anthropic, Google, or a self-hosted model. Here's how it works: the gateway fingerprints each prompt prefix, stores it centrally, and applies the right provider-specific caching mechanism automatically. When you switch providers or fail over during an outage, your cached context follows you. Early testing shows input token costs dropping by up to 70% on repeated system prompts and long documents, with latency improvements of 40% or more on cache hits. Tired of paying to send the same instructions thousands of times a day? Enable it with one configuration flag. Cache rules, TTLs, and per-team isolation are available in the dashboard today.

輸出 266 tok 延遲 3.8 s

是否遵守限制(字數上限、禁用詞清單、只能有一句問句)、文字風格的特徵,以及長度控制。

30 秒上手 Claude Sonnet 5.5

相容 OpenAI:換掉 base_url,SDK 不用換。POST /v1/chat/completions

from openai import OpenAI

client = OpenAI(
    base_url="https://synthorai.io/v1",
    api_key="sk-syn-...",
)

resp = client.chat.completions.create(
    model="claude-sonnet-5-5",
    messages=[{"role": "user", "content": "Summarize this diff"}],
    reasoning_effort="medium",
)
print(resp.choices[0].message.content)

關於 Claude Sonnet 5.5

  • 它提供 1M token 上下文視窗且無長上下文加價,最大輸出 128K,支援文字與圖像輸入。
  • 自適應思考預設開啟,預設強度為 high;最低檔 between_tools 會關閉「先想後答」,在 high 及以下強度可用,而把 thinking 設為 disabled 或手動指定 budget_tokens 都會回傳 400。
  • temperature、top_p、top_k 設為非預設值同樣回傳 400,強制工具呼叫也不受支援。
  • Anthropic 列出了五項會影響已在 Claude Sonnet 5 上執行程式碼的破壞性變更,工具呼叫之間的文字現在會以 thinking 區塊回傳。
  • 可快取提示詞的最小長度為 512 token。
  • Synthorai 透過與其餘模型相同的 OpenAI 相容 API 提供 Claude Sonnet 5.5。

常見問題

Claude Sonnet 5.5 API 可以免費試用嗎?

可以,新帳號有 10 次試用呼叫和最高 $1 的免費額度,免綁卡。以輸入價 $2/M 計算,光是這筆額度就夠你對 Claude Sonnet 5.5 發出約 62 次請求,每次約 8K token。

Claude Sonnet 5.5 最擅長什麼?

目前陣容中速度與智慧平衡最佳、與 Sonnet 5 同價:每百萬輸入 $2、輸出 $10、between_tools 可關閉「先想後答」。完整介紹請見「關於」一節,內容取自供應商官方的發布說明。

Claude Sonnet 5.5 的價格是多少?

在 Synthorai 上,Claude Sonnet 5.5 的輸入價是每百萬 token $2,輸出價是每百萬 token $10。這就是供應商官網價,平台不加價。快取命中的輸入 token 以 $0.1/M 計費。

Claude Sonnet 5.5 的上下文長度是多少?

Claude Sonnet 5.5 的上下文長度(上下文視窗)是 1M token。單次回覆最多輸出 128,000 個 token。

Claude Sonnet 5.5 支援提示詞快取(prompt caching)嗎?

支援,需主動開啟:以 cache_control 中斷點標記穩定前綴。快取命中的輸入 token 以 $0.1/M 計費(未命中為 $2/M);提示詞要有 512 個 token 以上的穩定前綴,才能命中快取(TTL 預設 5 分鐘,可選 1 小時)。 提示詞快取指南 →

如何開通 Claude Sonnet 5.5?

把現有 OpenAI SDK 的 base_url 指向 "https://synthorai.io/v1",model 設為 "claude-sonnet-5-5" 就可以了。一把 API 金鑰就能呼叫閘道上的所有模型。

Claude Sonnet 5.5 的知識截止日期是什麼時候?

Claude Sonnet 5.5 的知識截止日期為 2026-06,依據供應商官方文件(查核日期:2026-10-08)。

相關模型

比較

規格與廠商 benchmark 均附來源連結。閘道支援情況說明的是 Synthorai 的實際實作。封存範例是單次閘道請求,不代表模型的一般效能保證。價格來自公開價格介面或建置時使用的快照。

取得 API 金鑰 算算你的成本 →