Together Link: open models in the harness you already use. Start with one command today.
https://www.together.ai/blog/together-link-frontier-quality-open-models-in-the-harness-you-already-use📌 一行指令切換開源模型,Coding Agent 帳單砍半
TL;DR：Together Link 讓 Claude Code 等既有工具直接接上開源模型,省下一半以上費用。
工程團隊每月在 coding agent 上燒掉數萬到數百萬美元,原因往往不是任務難,而是每個任務——不管是改一行字還是整包重寫——都丟給同一個頂規閉源模型處理。Together Link 想解決的正是這種「用大砲打蚊子」的浪費。
🤔 開源模型已經追上來了
Coding agent 如今是工程團隊的標配,但不分任務難度都跑同一個 premium 模型,成本很快失控。Together AI 指出,像 Kimi K3、GLM 5.3 這類開源模型已經能處理許多最難的程式任務,價格只是閉源模型的一小部分;而 GLM 5.3 Flash、DeepSeek V4.1 Flash 這類輕量模型則足以應付日常的程式任務。
🧩 接進你本來就在用的工具,不用改習慣
Together Link 支援 Claude Code、Claude Desktop、Codex（ChatGPT app 與 CLI）、OpenCode 與 Pi。設定與登入方式維持原樣,不需要學新東西,要切回原本的閉源模型也只要一道指令。安裝方式是執行:
curl -fsSL https://link.together.ai/install | bash執行後 agent 開啟方式跟平常一樣,只是底層連到了 Together AI,使用的是你自己的 Together AI API key（需先註冊帳號取得）。
🧩 「Auto」模式:依任務難度自動路由
Together AI Router 提供「Auto」模式,讀取每個 session 的第一個任務內容,決定送去哪個模型:簡單修正交給快速、低成本的模型,困難問題則交給具備 frontier 能力的模型。如果你帶了自己的 Anthropic key,路由會在 Opus 5.5 與 GLM 5.3 之間選擇;沒有的話則在 GLM 5.3 與 GLM 5.3 Flash 之間選擇。路由判斷只在每個 session 開始時執行一次,因此 prompt caching 的效益不受影響。
📊 每個 session 都看得到省了多少
Together Link 提供 per-session 的費用追蹤器,直接對照「如果整段 session 跑在 Opus 5.5 上要花多少錢」。計費走你原有的 Together API key,可用 serverless pay-as-you-go 或 credit packs,不需要另外簽約。Together AI 表示,其 serverless 推論基礎設施就是開發者在 OpenRouter 上已經在選用的同一套,截至 2026 年 9 月 30 日,Together AI 在 OpenRouter 上承載 DeepSeek V4.1 Flash 40.8%、GLM 5.3 Flash 28.2%、Kimi K3 23.1% 的 token 流量份額。
⚠️ 省多少因任務組合而異
「省下 50% 以上」是 Together AI 自家的整體宣稱,實際節省幅度取決於團隊任務難度分布——簡單任務佔比越高,路由到輕量模型的效益自然越明顯。
🎯 實務啟示
如果你的工程團隊已經在用 Claude Code 或 Codex,且對逐月攀升的模型帳單有感,Together Link 的切換成本看起來很低:一道安裝指令、既有的工作流程不變、隨時可以切回原廠模型。值得先在小範圍團隊試跑幾個 session,實際比對費用追蹤器給出的節省數字再決定是否全面推行。
🔗 來源
- 標題：Together Link: open models in the harness you already use. Start with one command today.
- 作者／機構：Together AI
- 連結：https://www.together.ai/blog/together-link-frontier-quality-open-models-in-the-harness-you-already-use
#TogetherAI #OpenModels #CodingAgent #ClaudeCode #LLMRouting #CostOptimization #GLM #DeepSeek #KimiK3 #DevTools
原始資料 Together AI · 收集於 2026-10-05
摘要原文
Summary Together Link connects the harness your team already uses to the best open models on Together AI and cuts your spend by over 50%. Same agent, same workflow, a fraction of the bill. Coding agents are now commonplace for engineering teams, and every task you give them, from a one-line fix to a full rewrite, often runs on the same premium model. At scale that gets expensive quickly: engineering orgs are spending anywhere from tens of thousands to millions of dollars a month on closed models. Open models have meaningfully closed the gap between top models from closed labs. Leaders like Kimi K3 and GLM 5.3 now handle many of the hardest coding tasks at a fraction of the per-token price, while small and capable models like GLM 5.3 Flash and DeepSeek V4.1 Flash handle everyday coding tasks. Together Link brings them into the tools your team already knows. Works where you already work Together Link supports Claude Code, Claude Desktop, Codex (in the ChatGPT app and the CLI), OpenCode, and Pi. Your team keeps working normally: settings and logins stay as they were, there's nothing new to learn, and going back to the native closed models takes one command. You can be up and running on open models in just a few minutes. Getting started curl -fsSL https://link.together.ai/install | bash Run one command to set up Together Link: Your agent opens exactly as before, now connected to Together AI with your Together AI API key. Create an account for an API key. Use “Auto” to pick a model with the Together AI Router: It reads each session's first task and sends it to the right model: quick fixes go to fast, low-cost models, and hard problems get frontier capability. If you bring an Anthropic key, it routes between Opus 5.5 and GLM 5.3. Without one, it routes between GLM 5.3 and GLM 5.3 Flash. Routing happens once per session, so prompt caching keeps working. See your savings after each session. A per-session tracker shows what you spent next to what the same session would have cost on Opus 5.5. Billing runs on your existing Together API key, against serverless pay-as-you-go or credit packs. You don't need a separate contract. Built on the serverless platform developers already choose Together Link runs on Together's serverless inference, the same infrastructure developers already pick for these models on OpenRouter. Together AI serves the largest share of OpenRouter tokens for DeepSeek V4.1 Flash (40.8%), GLM 5.3 Flash (28.2%), and Kimi K3 (23.1%), with competitive speed and pricing across the board. ( As of 9/30/2026 footnote ) Next steps Create a Together AI account. Read our docs . Rolling it out across your engineering org? Talk to our team and we'll help you plan it.
由 claude-code/sonnet 自動生成