Krill, an AI relay service, launched a 618 promotion from June 15–18, 2026, reducing base Codex model rates to as low as 0.15 and offering a 66% discount coupon on Codex plans. With a 10-person group buy, the effective rate reaches 0.1 Chinese yuan per US dollar. Existing Codex plan holders on June 15 will have their quotas adjusted to the 0.1 level. Claude model access is discounted only via balance top-ups, not plans. The service uses Pro accounts and emphasizes cost transparency.
A V2EX user reported that a friend purchased a GLM annual subscription as a backup while primarily using OpenAI's Codex and ChatGPT. After recent policy-driven access restrictions (possible reference to “Fable” or similar incidents), that backup proved strategically valuable. The user warns against sole dependence on providers like OpenAI or Anthropic, whose policies can cut off access without notice, and plans to similarly secure a GLM annual plan. The post highlights growing community concerns over API dependency and the importance of having fallback options.
A V2EX user asks whether enterprises manually overclock GPUs when running local large language models (LLMs) in production, or if they simply use them at stock settings like CPUs. The post contains only the question, with no answer or additional information provided.
A user set up a GPT Pro subscription relay for colleagues using sub2api on a US-based CN2 server with a ping latency of approximately 160ms. The relay exhibits high time-to-first-token (TTFT), making responses slow. The user is seeking optimization advice without clear direction.
A Chinese developer revived their WeChat mini-program '旅泡泡' to plan a family trip to Hong Kong. The app lets users input destination, dates, budget, and preferences (e.g., 'traveling with children, don't make it too tiring'), then uses models like DeepSeek to interactively generate a draft itinerary. It is positioned as a 0-to-1 framework generator rather than a replacement for detailed guides. Generated plans can be saved and shared with friends via WeChat cards. The developer is now seeking community feedback on whether users prefer chat-based or form-based interaction, and if itineraries should be hourly or high-level.
A new AI service called Claude-fable-5 from a source factory has been launched. The service is accessible at muskapi.cc. Additionally, users can receive free credits for GPT-5.5 through a separate link. This promotion post on v2ex highlights the availability of these AI offerings. The author aims to attract users to test the new API services.