Top stories today:

1. GPT-4.1 coding-focused models released to take on Google, Anthropic
2. FTC argues Meta holds monopoly over “personal social networking”
3. Trump admin probes pharma, chip sectors in bid to impose tariffs
4. China said to stop Boeing jet deliveries by airlines amid U.S. trade war
5. GM faces test to deliver on $35B EV push under Trump-era pressures

0. Data and calendar

Listen to our AI-generated podcast summarizing today’s newsletter (beware of hallucinations):

1. GPT-4.1 coding-focused AI models released by OpenAI to take on Google, Anthropic

GPT-4.1 has scored 52.4% on the Aider Polyglot Coding Benchmark, trailing behind DeepSeek V3 and Claude 3.7 Sonnet:

• GPT -4.1 mini and GPT-4.1 nano are the other 2 models in the family.

• GPT-4.1 family supports up to 1M tokens of context.

• Available only through the API and not via ChatGPT.

• GPT-4.1 costs $2/M input and $8/M output tokens, vs. GPT-4o’s $2.50/M input and $10/M output tokens.

• GPT-4.1 mini is priced at $0.40/M input and $1.60/M output tokens and GPT-4.1 nano at $0.10/M and $0.40/M.

• GPT-4.5 will be phased out from OpenAI’s API on Jul. 14.

The GPT-4.1 family is faster than the GPT-4o series

GPT-4.1 is also providing more accurate results on the SWE-bench Verified than GPT-4o, per OpenAI

GPT-4.1 has scored 72% on MCQs from 30-60 min. videos without subtitles, per OpenAI

• Audio support appears to be missing, which was available on GPT-4o.

• GPT-4.1 will become less reliable as more input tokens are provided.

• Relatedly, o3 and o4-mini, rumored to excel at proposing new experiments, may debut soon.

3. FTC argues Meta holds monopoly over “personal social networking” market