DeepSeek Harness plugin ecosystem is booming as cheap, mid-tier-but-solid model draws tinkerers
op7418 · x · 2026-10-27
op7418 tried several popular DeepSeek Harness plugins and found the ecosystem thriving. Screenshot shows a full session dashboard: plugin marketplace, context stats (60.8k/1M tokens, 97.15% cache hit), token/latency breakdown (tool execution 53.2% of a 5m33s session) and per-call cost (1.5M tokens / ¥0.32).
His take: DeepSeek is extremely cheap yet solidly mid-tier-strong, so many users—especially young tinkerers, likened to the old phone-flashing/ROM crowd—now run everything on DeepSeek Harness.
Related event: DeepSeek Harness Plugin Ecosystem Booms as Creator Lists Top Five Plugins(2 posts)→
More from coding & agent
- Teaching Grok Bot to take your work: five ways to hand off tasks, from drafting to cross-app workflows — Scobleizer · 2026-10-27
- AutoHarness: open-source self-learning skill layer for Claude Code lifts CORE-Bench 42%→78% — thisguyknowsai · 2026-10-27
- GenXR-Bench: AI coding agents fail at XR apps, best model scores just 50.9% — rohanpaul_ai · 2026-10-27
- Alibaba open-sources OpenCodeReview, an AI code reviewer claiming 9x fewer tokens than Claude Code — thisdudelikesAI · 2026-10-27
- Google's Thinking-in-Bets Nearly Eliminates False Reversals in LLM Agents — google · 2026-10-27
- 4,165 Agent Turns Later: Small Open-Weight Models Struggle at Real Coding Work — Vaibhav Jain · 2026-10-27