# Together Link puts open models behind the coding-agent harnesses developers already use Together AI's new beta keeps the coding-agent harness while swapping the model/provider layer underneath it, making model choice less tightly coupled to Claude Code, Codex, OpenCode or Pi. Together Link connects six existing coding-agent/desktop harnesses to open models with reversible profiles, per-session routing and cost receipts. The important shift is portability at the harness boundary, not Together's unverified savings claim. - Status: Active - Published: 2026-10-06T06:31:12+13:00 - Updated: 2026-10-06T06:31:12+13:00 - Categories: Artificial Intelligence, Web Development, SaaS, AI SaaS, AI Coding, Developer Tools - Tags: Coding agents, model routing, open models, Together AI - Canonical HTML: https://beyondthe.news/dossiers/together-link-open-models-claude-code-codex-opencode-pi ## What changed Together AI launched Together Link in beta on October 5. The free, MIT-licensed tool connects Claude Code, Claude Desktop, ChatGPT Desktop, Codex CLI, OpenCode and Pi to models served by Together AI while preserving the user's existing harness, settings and normal provider setup. Terminal integrations use temporary/per-session configuration; desktop integrations use reversible profiles. The default Auto Router chooses a model once per session, preserving prompt caching, and each proxied session can report token and dollar totals. ## Why it matters Coding-agent products increasingly bundle two distinct layers: the harness that manages tools/context/workflow and the model that reasons. Together Link makes that boundary operationally visible by letting developers keep a familiar harness while changing the model/provider underneath it. That can make cost and model experiments easier without retraining a team on another agent interface. The economics need caution: Together says open-model use can cut agent spend by more than half or 50–80%, but that is a vendor claim and cost per token is not the same as cost per successfully completed task. ## The harness stays; the model layer moves Together Link supports Claude Code, Claude Desktop, ChatGPT Desktop, Codex CLI, OpenCode and Pi. Rather than asking teams to adopt another coding agent, it adapts those existing interfaces to Together-served models. ## Configuration is deliberately reversible For terminal agents, Together says configuration is applied only for the launched session. Desktop integrations use separate profiles that can be turned off, reducing the switching cost of testing another model provider. ## Routing happens once per session Auto Router examines the initial session task and chooses a model for that session rather than rerouting every turn. Together says this preserves prompt caching. With an optional Anthropic API key it can also route harder work to Opus 5.5; without one it stays on Together-served models. ## Cost claims need task-level evidence Together markets savings of more than 50%, and its product page says 50–80% versus running every session on Opus 5.5. Those figures are vendor-produced. Independent harness research reinforces that the harness can materially affect results and cost, which is precisely why token-price comparisons alone cannot establish cost per solved task. ## Key details - Together Link launched in beta on October 5, 2026. - It supports Claude Code, Claude Desktop, ChatGPT Desktop, Codex CLI, OpenCode and Pi. - Together Link itself is free and MIT licensed; model usage is billed through a Together AI API key. - Terminal integrations use temporary session configuration while desktop apps use reversible profiles. - Auto Router selects a model once per session rather than rerouting each turn. - Each proxied session can print token and dollar totals and the usage command can report recent spend. - Together's 50–80% savings claim is vendor-produced and is not independently established as cost per completed coding task. ## Builder takeaways - Separate your evaluation of a coding harness from your evaluation of the model behind it; Together Link makes it practical to hold the interface relatively constant while changing models. - Compare cost per accepted task or merged change, not merely per-token prices or the session receipt. - Use the reversible profile/session boundary for controlled trials before changing a team's default provider. - Check tool compatibility and output quality on your own repositories because an open model running inside a familiar harness is not automatically equivalent to the harness vendor's native model. ## What to watch - Whether Together Link expands beyond macOS/Linux or adds more coding-agent harnesses. - Independent same-harness comparisons of Together-served open models against native Claude/Codex pairings. - Whether routing evolves beyond one model decision per session. - Whether closed-harness vendors change compatibility behaviour in response to provider-swapping layers. ## Uncertainties - Together's savings and frontier-quality claims are vendor claims and have not been independently verified for Together Link. - The product is beta and supported harness behaviour may change. - Independent harness studies show harness choice can affect success and cost, but they do not directly benchmark Together Link. ## Sources - [Together Link: frontier-quality open models in the harness you already use](https://www.together.ai/blog/together-link-frontier-quality-open-models-in-the-harness-you-already-use) — Together AI · primary · 2026-10-05T00:00:00+13:00. Launch announcement, supported harnesses, routing and vendor savings claims. - [Together Link](https://www.together.ai/link) — Together AI · primary documentation · 2026-10-05T00:00:00+13:00. Current beta behaviour, supported integrations, pricing examples, reversible configuration and usage reporting. - [Harness or Model? Isolating the Harness Effect in Agentic Coding with a Contamination-Controlled Private Suite](https://arxiv.org/abs/2609.11987) — arXiv · independent research · 2026-09-08T00:00:00+12:00. Independent context showing that harness choice can affect coding-agent correctness and cost; does not benchmark Together Link.