6 articles tagged
Devin Desktop v3.8.20 lets you stop sharing terminal activity and local edits with third-party ACP agents, but both settings default to on and neither appears in the Agent Client Protocol documentation.
Poolside's 118B open-weight coding model is 100x cheaper on input than Claude Fable 5 and scores 70.2% to Fable's 88% on Terminal-Bench 2.1. Here is the routing decision that follows from those two numbers.
DeepSeek's V4-Flash-0731 (July 31, 2026) adds native Responses API and an official OpenAI Codex path at $0.14/$0.28 per million tokens. How to wire a coding agent to it, its agent-benchmark scores, and the cost math.
Anthropic shipped Claude Opus 5 on 24 July 2026 and OpenAI shipped GPT-5.6 (Sol, Terra, Luna) on 9 July. For coding agents the two flagships cost almost the same on input and diverge on output, benchmarks and token
Adversa AI tested 11 open-source coding agents in June 2026 and beat the safety filter on 10 of them with shell tricks that are 30 years old. How the bypass works, which designs held, what to change now.
Z.ai's GLM-5.2 is the strongest open-weight coding model of 2026 and a third of Claude Opus 4.8's input price. Self-hosting it is another bill: 753B parameters, 1.5 TB of weights, a GPU node rented by the hour.
