****OpenAI DoD/Pentagon AWS Deal, MSFT Rift, Musk Lawsuit Escalation, Anti-China Model Coop****
Key Questions
What is GPT-5.4 Pro's performance in recent benchmarks?
GPT-5.4 Pro tops benchmarks and saw an 8.9% usage increase after OpenClaw was banned in Claude subscriptions. It outperforms models like GPT-4o-mini.
How does GPT-5 medium compare to Mistral Small 4?
Mistral Small 4 (Reasoning) is compared to GPT-5 medium in benchmarks. Specifics highlight competitive reasoning capabilities between the models.
Why was Sora shut down by OpenAI?
OpenAI closed its Sora video application six months after launch, described as a '$15M/day mistake.' It followed user losses of 2.5 million.
What are the rumors about OpenAI's Spud model?
Spud, potentially GPT-5.5, is rumored for Q2 release with intuitive multimodal and agent-like behavior. It's positioned as a major upcoming advancement.
What is OpenAI's Safety Fellowship program?
The OpenAI Safety Fellowship supports independent AI safety research, offering up to $15,000 in monthly compute. It's a pilot for external researchers announced in April 2026.
How is Anthropic pressuring OpenAI financially?
Anthropic reached $30B ARR in April 2026, surpassing OpenAI, boosted by a Meta contract. This intensifies competition in AI revenue.
What new integrations is ChatGPT developing for its superapp?
ChatGPT is integrating services like DoorDash, Spotify, and Uber as part of its superapp general availability. Sam Altman noted challenges like adding timer functions may take another year.
What voice and other limits are mentioned for upcoming GPT models?
Discussions include voice limits in models like GPT-5 series. Benchmarks and comparisons also cover medium models versus rivals like Mistral.
DoD deals via AWS (post-Anthropic refusal); MS rift; Musk escalates lawsuit demanding Altman/Brockman ouster ($134B damages, Apr27 trial); Frontier Forum blocks Chinese theft. Natsec/legal tensions peak.