OpenAI Codex July 27: Fourth Outage in Four Days Throws 503 biscuit_baker_service_me_circuit_open, Codex CLI 0.145.0 /import Migrates Cursor and Claude Code Settings, ChatGPT Voice Now Drives Codex Agents

OpenAI outage: ChatGPT, API and Codex CLI down with 503 biscuit_baker_service_me_circuit_open errors, fourth incident in four days

• On July 25 ChatGPT, the OpenAI API and Codex CLI went down together, returning 503 biscuit_baker_service_me_circuit_open errors, OpenAI's fourth incident in four days.

Editor’s read: Four incidents in four days is a reliability pattern, not bad luck. Wire retry with backoff and a fallback model into anything production.

Claude Opus 5 vs GPT-5.6 Sol benchmarks: Opus 5 wins SWE-bench Pro 79.2% vs 64.6%, Sol Ultra takes Terminal-Bench at 91.9%

• On July 25 CodingFleet published head-to-head benchmarks: Claude Opus 5 wins SWE-bench Pro at 79.2% versus 64.6% for GPT-5.6 Sol, while Sol Ultra takes Terminal-Bench at 91.9%.

Editor’s read: Sol Ultra still owns terminal-heavy work. Pick the model per bench that matches your workload, not the headline number.

OpenAI Codex CLI v0.145.0: /import migrates Cursor and Claude Code settings, multi-agent V2 stabilized, audio inputs, Bedrock login

• Released July 21: a new /import command migrates your Cursor and Claude Code settings into Codex in one step, multi-agent V2 is stabilized, and the CLI gains audio inputs and AWS Bedrock login.

Editor’s read: /import is OpenAI openly courting Cursor and Claude Code switchers. Migration friction just stopped being an excuse to avoid a trial run.

OpenAI Codex goes hands-free: ChatGPT Voice lands on desktop powered by GPT-Live and can drive Codex agents, as Codex passes 8 million users

• On July 23 an OpenAI engineer teased a Codex reveal for July 24 as Codex passed 8 million users.
• On July 24 the announcement landed: ChatGPT Voice arrives on the desktop app, powered by GPT-Live, and can drive Codex agents hands-free.

Editor’s read: Voice-driven agents sound like a demo gimmick until you try dictating a refactor while reading the diff. Worth one honest attempt before judging.

Nader Dabit: SWE-1.7 Lightning at 1,000 tok/s is 1,200% faster, while Grok 4.5 and Claude Opus 5 sit alone on the Pareto frontier

• On July 25 Nader Dabit reported SWE-1.7 Lightning running at 1,000 tok/s, 1,200% faster with little intelligence tradeoff, and placed Grok 4.5 and Claude Opus 5 alone on the price-intelligence Pareto frontier.

Greg Isenberg: 20 reasons this is the best time to build, from $20/mo superintelligence to LLM search minting the next millionaires

• On July 21 Greg Isenberg listed 20 reasons this is the best time to build: AI agents outnumbering humans, $20/mo superintelligence, and LLM search minting the next wave of millionaires.

levelsio: AI coding equalized everyone, apps that took a year now take hours, and speed is the last differentiator

• On July 20 levelsio argued AI coding has equalized developers: apps that took a year now take hours, speed is the last differentiator, and only AI companies are winning.