DeepSeek Harness preset switching limitation: mid-session changes break model comparison experiments
Physical-Neat-5211 · reddit · 2026-08-14
The author discovered that in DeepSeek Harness, once a session has messages, the Web UI prevents switching presets. This is because a preset changes not only the model but also tools, approval rules, system prompt, and agent loop, so switching mid-session would leave old tool calls inconsistent with the new runtime, ruining comparison experiments. The author uses an OpenAI-compatible API and ZenMux gateway, starting a fresh session for each model while keeping presets unchanged, and verifies routing via Harness trace and gateway logs. Harness is still a developer preview, so behavior may change.
More from coding & agent
- Terminal Night 3 talk slides released, missing demo video to follow — 4310sy · 2026-10-16
- Agent developers are sitting on unused API credits that can exceed actual usage — EnvironmentalDay8864 · 2026-10-16
- Notion's bet on agents: abstract away deployment so users and agents just write code — simonguozirui · 2026-10-16
- Four Gemini 4 Argon agents beat Minecraft's Ender Dragon autonomously in 27 hours — Maxence Faldor · 2026-10-16
- Tencent's Terminal-RSI builds recursively self-improving terminal agents, hitting 19.84% on Terminal-Bench 4.0 — tencent · 2026-10-16
- Tsinghua's AutoSFT uses multi-agent collaboration to automate SFT data research — TsinghuaNLP · 2026-10-16