GPT-6.1 Sol’s default context in Codex increased to 500k
要闻
Tibo, OpenAI’s Codex lead, said GPT-6.1 Sol’s default context in Codex and Work has increased from 272k to 500k; its maximum remains 1M, and long-context usage carries no extra subscription-quota charge.
Tibo, OpenAI’s Codex lead, said the default context length for GPT-6.1 Sol in Codex and Work has been changed from 272k to 500k. According to him, the roughly 80% increase is intended to reduce compaction and support tasks that need to retain large amounts of context and detail for extended periods. He said the value was tested and is suitable for most users. Users can still raise or lower the context length themselves, while GPT-6.1 Sol supports a maximum context length of 1M. After this change, GPT-6.1 Sol does not charge a long-context fee when usage is deducted from a subscription quota.
Tibo surveys Pro users on usage preferences: double the weekly allowance or remove the 5-hour limit
开发生态
Tibo, head of OpenAI Codex, opened a public poll for ChatGPT Pro users on two quota options: doubling weekly usage while keeping the five-hour limit, or removing that limit while keeping the current weekly total.
Tibo, head of OpenAI Codex, launched a public poll asking ChatGPT Pro subscribers which of two quota changes they prefer. The first option keeps the existing five-hour usage limit and doubles the total amount of usage allowed per week. The second removes the five-hour frequency limit while leaving the current weekly usage total unchanged.
Claude grants select paid users $100 in initial Projects credits
产品应用
Claude has recently granted some paid users who were accepted into the Projects test up to $100 in initial credit. The credit is limited to Projects, does not count toward the regular usage limit, and remains valid for 24 hours after setup.
Claude recently provided subscribers whose applications for the Projects test were accepted with up to $100 in initial Projects usage credit. When a user creates a new project from the sidebar, the account automatically receives the credit. According to user reports and prompts in the official interface, the credit can be used only within Projects and does not count toward the account’s regular usage limit for its standard billing cycle.
The initial credit is valid for 24 hours after setup is completed. The source does not specify which subscription plans are eligible, the total amount being distributed, or the scope of the test application process.
Musk showcases SpaceX’s AI division renamed SpaceXSI
行业动态
A screenshot posted by Elon Musk on X shows that SpaceX’s artificial intelligence division has changed its account ID to “SpaceXSI,” renamed itself “SpaceX Super Intelligence,” and adopted a gold business verification badge, putting the new name into official use.
SpaceX’s artificial intelligence division now uses SpaceXSI as its account ID on X and displays the name SpaceX Super Intelligence. A screenshot of the page posted by Elon Musk on X shows that the account has official verification and a gold business verification badge. Based on the information shown in the screenshot, the name change has taken effect on X.
Google begins limited testing of its new model Gemini 4 Argon
前瞻与传闻
According to user reports, Google has begun gray testing its next flagship model, Gemini 4 Argon, on the Gemini web app and Antigravity, allowing some users to try its reasoning and coding capabilities without an API.
According to user reports, Google has started gray testing Gemini 4 Argon on the Gemini web app and Antigravity. Some selected users can call the model directly through the existing gemini model options without using a separate API.\n\nThe test provides access to deep reasoning and coding capabilities, but it remains under strict rate limits. The range of eligible users and available usage quotas are restricted. The information comes from user reports, and the source text contains no statement from Google officially explaining the test.
