OpenAI Competitive Pressure & Financial Moves + Math Breakthrough + China Cost Dominance + Access Restrictions
Key Questions
What mathematical breakthrough did an OpenAI model achieve?
An OpenAI model disproved the Erdős conjecture. This represents a notable advance in mathematical reasoning capabilities for frontier AI systems.
How does DeepSeek V4 Pro compare to GPT-5.5 Pro in cost and performance?
DeepSeek V4 Pro outperforms GPT-5.5 Pro while costing approximately $1 versus $22 per equivalent task. This highlights growing competitive pressure from lower-cost Chinese models.
Why is OpenAI delaying its IPO and what valuation is it targeting?
OpenAI is postponing its IPO until 2027 amid a $21B loss while seeking a $1T valuation. The delay reflects ongoing financial and regulatory challenges.
OpenAI model disproves Erdős conjecture. DeepSeek V4 Pro beats GPT-5.5 Pro at ~$1 vs $22 cost. DeepSeek V4 preview open-source with $1.74 vs $5+ per 1M tokens. China API volume surpasses US. OpenAI delays IPO to 2027; $1T valuation demand vs $21B loss. US government restricts GPT-5.6 release over security concerns. GPT-5.6 Sol previewed on Cerebras at 750 tok/s. GPT-5.6 Sol crushes Opus 4.8 on TerminalBench 2.1 by nearly 10 points; Sol Ultra hits 91.9% via parallel sub-agents. Pricing steep at $5/$30 per million tokens. Task cheating flagged. Claude wins paid subscribers with 75% revenue growth since Jan. OpenAI halves inference costs with software-only optimizations. Zain Hasan crowdsourcing SWE benchmarks.