Enable GPT-6.1 Sol BYOK and set agent.max_idle_retained_threads in Zed
Zed v1.23.1-pre adds BYOK for GPT-6.1 Sol with an OpenAI API key, agent.max_idle_retained_threads to unload idle agent threads, JSONL/NDJSON table previews via cmd-k v / ctrl-k v, automatic Google AI overload retries, and a fix so OpenAI MCP tool calls no longer require optional fields.Why this loop
This prerelease changes how you run agents in Zed, not just the editor chrome. GPT-6.1 Sol is now available via BYOK with an OpenAI API key, so the Agent Panel can use Sol with keys you already hold. Set agent.max_idle_retained_threads to cap how many idle agent threads stay loaded; the same release also unloads external-agent sessions that remained loaded after their conversation closed during loading, so long agent sessions leak less RAM. Preview JSONL and NDJSON as tables with cmd-k v (macOS) or ctrl-k v (Linux/Windows) to inspect eval traces, agent logs, and NDJSON telemetry beside the source. Two correctness fixes unblock existing agent work: OpenAI tool calls to MCP servers no longer treat optional fields as required, and Google AI retries when a model is overloaded with clearer errors. OpenCode model resolution is faster and deprecated models are hidden. If Zed Cloud failed with UnknownIssuer behind a corporate TLS-inspecting proxy, it should connect when that proxy root CA is in the system trust store.
Proposed actions
- In the Zed Agent Panel, add your OpenAI API key and select GPT-6.1 Sol as the model (BYOK added in v1.23.1-pre).
- Add this key to ~/.config/zed/settings.json: "agent.max_idle_retained_threads": 2 so idle agent threads—including external-agent sessions closed while still loading—are not kept in memory indefinitely.
- Open a .jsonl or .ndjson eval/log file and press cmd-k v (macOS) or ctrl-k v (Linux/Windows) to open a tabular preview alongside the source.
- Re-send any OpenAI agent turns that failed on MCP tools because optional schema fields were treated as required; that mismatch is fixed in this release.
- If Google AI previously failed with overload errors, switch those threads back to the Google model—Zed now retries automatically. If Zed Cloud failed with UnknownIssuer, install the corporate TLS-inspecting proxy root CA in the system trust store and reconnect.
Agent prompt
Paste into your agent or query via MCP (get_agent_prompt) — free, no extra AI costAGENTS.md / CLAUDE.md / GEMINI.md rule update
DevAgentRadar → Repo Harness Rule Patch
Goal: update our repository's permanent harness instructions based on this release.
Update the file that tool actually reads. Do not dump everything into one file.
- Claude Code → CLAUDE.md plus
.claude/(skills, hooks, settings, agents, commands). It does not read AGENTS.md natively; start CLAUDE.md with@AGENTS.md. - Codex, Copilot, Cursor, Factory Droid, Grok Build, Roo Code, Goose, OpenCode, Amp, Zed, Aider → AGENTS.md.
- Gemini CLI / Antigravity → GEMINI.md. AGENTS.md only if
context.fileNameis set. - Cursor glob-scoped rules →
.cursor/rules/*.mdc(plain.mdis ignored), not a second constitution. - Codex MCP →
.codex/config.toml, not.mcp.json. - Copilot extra instructions →
.github/copilot-instructions.md; custom agents →.github/agents/. - Windsurf →
.windsurf/rules(do not assume AGENTS.md). - Cline →
.clinerules. - Procedures → a skill (
SKILL.md). Enforcement the model must not skip → a hook. Isolated roles → subagents. - Do not fork the same rule into three tool files.
Context
Assistant: Zed AI Proposal: Enable GPT-6.1 Sol BYOK and set agent.max_idle_retained_threads in Zed Summary: Zed v1.23.1-pre adds BYOK for GPT-6.1 Sol with an OpenAI API key, agent.max_idle_retained_threads to unload idle agent threads, JSONL/NDJSON table previews via cmd-k v / ctrl-k v, automatic Google AI overload retries, and a fix so OpenAI MCP tool calls no longer require optional fields. Primary source: https://github.com/zed-industries/zed/releases/tag/v1.23.1-pre
Why it matters
This prerelease changes how you run agents in Zed, not just the editor chrome. GPT-6.1 Sol is now available via BYOK with an OpenAI API key, so the Agent Panel can use Sol with keys you already hold. Set agent.max_idle_retained_threads to cap how many idle agent threads stay loaded; the same release also unloads external-agent sessions that remained loaded after their conversation closed during loading, so long agent sessions leak less RAM. Preview JSONL and NDJSON as tables with cmd-k v (macOS) or ctrl-k v (Linux/Windows) to inspect eval traces, agent logs, and NDJSON telemetry beside the source. Two correctness fixes unblock existing agent work: OpenAI tool calls to MCP servers no longer treat optional fields as required, and Google AI retries when a model is overloaded with clearer errors. OpenCode model resolution is faster and deprecated models are hidden. If Zed Cloud failed with UnknownIssuer behind a corporate TLS-inspecting proxy, it should connect when that proxy root CA is in the system trust store.
Suggested actions
- In the Zed Agent Panel, add your OpenAI API key and select GPT-6.1 Sol as the model (BYOK added in v1.23.1-pre).
- Add this key to ~/.config/zed/settings.json: "agent.max_idle_retained_threads": 2 so idle agent threads—including external-agent sessions closed while still loading—are not kept in memory indefinitely.
- Open a .jsonl or .ndjson eval/log file and press cmd-k v (macOS) or ctrl-k v (Linux/Windows) to open a tabular preview alongside the source.
- Re-send any OpenAI agent turns that failed on MCP tools because optional schema fields were treated as required; that mismatch is fixed in this release.
- If Google AI previously failed with overload errors, switch those threads back to the Google model—Zed now retries automatically. If Zed Cloud failed with UnknownIssuer, install the corporate TLS-inspecting proxy root CA in the system trust store and reconnect.
Config surfaces this release may change
- MCP servers — check your repo before applying
After you finish
Do not report this as applied to DevAgentRadar. You cannot write the visitor's loop.
Tell the human: open https://devagentradar.com/proposals/zed-ai-v1-23-1-pre-enable-gpt-6-1-sol-byok-and-set-agent-max-idle-retain and mark Applied, Skipped, or Failed. Proposal id: 610b9f92-3174-45db-aa4f-8647f54a7c2c
Instructions:
- Read the existing context files and settings for the tools in this repo.
- Draft an explicit Git diff. Update ONLY agent harness rules or tool configuration.
- Call out deprecated flags, obsolete habits, or changed permission boundaries.
- Do not touch application logic.
Start with the proposed diff and say which layer and which tool's file it belongs in.
Your loop
This browser · no sign-in · not shared as “you”After you run the prompt
Only you can mark this. Agents cannot write your loop.
Your decision stays on this device. A public tally appears after a few votes.