OpenAI Discovers Additional Consumption Issues with Codex, Full Reset Tomorrow
OpenAI has identified the reasons for the rapid consumption of Codex quotas and found three areas of additional consumption: the inefficiency of repeatedly compressing context in long conversations with images; excessive usage of the Computer History feature in high consumption scenarios; and the automatic generation of conversation titles exceeding expectations. The person in charge, Tibo Sottiaux, previously stated that no overall anomalies were found, and some users were directed to sub2api and subscription sharing, but later admitted that some users experienced a decrease in cache hit rates, leading to accelerated quota consumption. The investigation pinpointed issues within Codex itself. Context compression is used for long tasks, and the current process incurs additional waste during multiple compressions. The Computer History feature can bring selected app and web operation records from Mac into ChatGPT and Codex. OpenAI will push a fix on Sunday, U.S. time, and perform a complete reset of Codex usage for all paid subscriptions, expected to go live at 2 PM Pacific Time, corresponding to 5 AM Beijing time on August 24. The team has also found a new optimization plan, which is expected to be advanced next week.
-- Price
This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.
You may also like

Citrini Research Claims Treasury-Fed Agreement Could Drive Up 30-Year U.S. Bonds

New Zealand ACT party proposes tax exemption for long-held crypto gains

Glassnode: 1.05 Million BTC Long-Term Holders in the $83,000 to $86,000 Range

ATF Confirms Major Incident Following Qilin's Appearance on Its Leak Site

Stable Sea adds 2 WisdomTree funds for corporate cash

Goldman Sachs Expects Nvidia's Revenue to Grow by 70% in Fiscal Year 2028

Sweden Allocates 270 Million Krona to Aid Ukraine

Bitcoin Holds Steady at $80,000 as Traders Focus on Walsh's Speech

Avalanche proposes zero-inflation validator reward model

Apple Plans to Ship Over 10 Million Foldable iPhones in the First Year

10-Year Won Interest Rate Swap Decreases by 5.50bp

California Pension Fund Increases Investment in SpaceX to $2 Billion

Mr&强|买美股上 WEEX Analyzes US Stocks and Cryptocurrency Market

Variable Robotics Releases WALL-SS Self-Regressive World Model

Jackson Hole Conference Focuses on Long-Term Policy Direction of the Federal Reserve

Deutsche Bank Predicts More Aggressive Debt Management Measures by U.S. Treasury

Concerns Over Dollar Depreciation Support Gold Prices as Market Awaits Walsh's Speech

Yili Hua Bullish, Plans to Close Long Positions at High Levels and Continue Buying

FinTax Completes $40 Million Seed Round Financing Led by YZi Labs

Jack Yi Predicts Bull Market Trend Will Last Until 2028, Advises Against Short Selling

SoftBank Negotiates to Acquire Majority Stake in 1X for $6 Billion

NVIDIA Qwen3.8-Flash-Next Achieves Over 16,000 Tokens/Second Throughput on GB300 NVL72

Copper Prices Surpass 90,000 Yuan/Ton, Driven by AI and Power Grid Demand

Miner Distribution at 51.2%, Airdrop at 20.4% Announced

European Central Bank's Pontes Distributed Ledger Project Set to Launch on September 21, 2026

Ed Yardeni Raises S&P 500 Year-End Target to 8400 Points, Cautious on AI Market

MiniMax M3Pro Parameter Scale Expected to Increase to Approximately 3T

Luis Caputo announces 2 trillion plan for mortgage loans with ANSES funds

Gold prices rise to $4,700.6 as central banks return to buying






