On Oct 5, Codex lead Thibault "Tibo" Sottiaux committed that for the next 28 days OpenAI will ship one clear improvement for most Codex/ChatGPT Work users each day, or give everyone a full usage reset. Day 1 is an inference speedup: GPT-6 Astra and GPT-6.1 Sol are ~50% faster by default on subscriptions (roughly 30 to 50 tokens/s), including partner tools using Sign in with ChatGPT. No API speed change was announced.

Key Takeaways

  • ✓28-day pledge: each day either one clear improvement or a full usage reset for everyone (OpenAI Developer Community thread)
  • ✓Day 1: GPT-6 Astra / GPT-6.1 Sol default speed ~+50%; Tibo cited roughly 30 to 50 tokens/s (closer to +67% if exact; both called approximate)
  • ✓Scope: OpenAI products plus every Sign in with ChatGPT partner (OpenCode, Pi, Amp, Devin), no config change, live within ~2 hours
  • ✓Subscription path only; no API speed change announced, so API-billed evals are unaffected
  • ✓One community user reported slower 6.1 Sol runs on a private benchmark after the change; unverified single data point
⚡

Key Decision Metrics at a Glance

TL;DR Verdict
🏆 Benchmark Rank72s read
Day 1: GPT-6 Astra / GPT-6.1 Sol default speed +50%; Tibo cited roughly 30 to 50 tokens/s (closer to +67% if exact; both called approximate)
Full Leaderboard→
💰 Cost & PlansCompare
in: $10 / out: $50
Compare Plans↗
🛠️ Ecosystem & IDERules
Codex · OpenCode · Amp
Setup IDE Rules↗
🧭

Turn your technical choice into a development budget

Compare 40 dev plans & simulate token costs vs $20/mo subscriptions

⚔️ Compare Model→💰 Compare plans & limits↗
🔗

Project Links & Resources

Direct Access
Direct access to official project resources and documentation
🌐
OpenAI Developer Community: Day 1 of 28 days of Quality of Life improvements or a full Reset
community.openai.com/t/day-1-of-28-days-of-quality-of-life-improvements-or-a-full-reset/1403525
↗
📝
Runtime Wire: OpenAI says GPT-6 Astra and Sol now stream 50% faster
runtimewire.com/article/openai-gpt-6-models-speed-optimization-tibo-sottiaux
↗
📖
Codex changelog
developers.openai.com/codex/changelog
↗
📝
OpenAI: Introducing GPT-6 Sol and Luna
openai.com/index/introducing-gpt-6-sol-and-luna/
↗
🔬

In-Depth Technical Analysis

Codex lead Thibault Sottiaux announced a 28-day pledge on Oct 5: each day OpenAI will either ship one clear improvement relevant to most Codex/ChatGPT Work users or grant everyone a full usage reset. Day 1 is a server-side inference optimization (OpenAI's Vaibhav Srivastav called it an inference improvement; details undisclosed) that makes GPT-6 Astra and GPT-6.1 Sol about 50% faster by default when used through a ChatGPT subscription, roughly 30 to 50 tokens per second. It applies to OpenAI's own clients and to partner tools that use Sign in with ChatGPT, including OpenCode, Pi, Amp and Devin, with no settings change and rollout within about two hours. No change to API speed was announced. The figures are approximate and describe streaming throughput, not end-to-end task time; one community user reported slower 6.1 Sol runs on a private benchmark, which is unverified. Developers can simply keep using gpt-6-astra or gpt-6.1-sol through their ChatGPT login and compare wall-clock time and tokens on a fixed task.

Action HubReady to adopt this in production?

Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.

⚔️ Launch VS Compare→💳 View Plans↗