🚀 OpenAI launches GPT-5.4 with Pro and Thinking models

OpenAI has launched GPT-5.4, its newest model built for professional work. With versions for everyday use, deeper reasoning, and high-performance tasks, it aims to deliver bigger context and fewer mistakes. One standout feature: the “Thinking” version can show its plan upfront, letting users guide the response as it unfolds. Let’s dive in!

In today’s insights:

LATEST DEVELOPMENTS

Evolving AI: OpenAI introduces GPT-5.4 with larger context and fewer errors.

OpenAI released GPT-5.4, its newest frontier model aimed at professional tasks. The model comes in three variants: the base version, GPT-5.4 Thinking for deeper reasoning tasks, and GPT-5.4 Pro for high performance workloads. The API now supports context windows up to 1 million tokens. Benchmarks show stronger results across computer use tests like OSWorld and WebArena and higher scores on knowledge work evaluations. The model produces fewer factual errors and introduces Tool Search, letting systems retrieve tool definitions only when needed.

GPT-5.4 Thinking can now share an upfront plan of its reasoning, letting users adjust direction mid-response and steer the output without needing extra prompts:

This pushes AI further from chat toy to actual work layer. OpenAI says GPT-5.4 cuts factual mistakes versus GPT-5.2, stretches context to 1 million tokens, and is getting better at spreadsheets, presentations, documents, and tool use, which means teams can hand off bigger chunks of real work instead of breaking everything into tiny prompts. Mercor’s benchmark result adds to that picture in law and finance, where the value is less about flashy demos and more about getting a draft deck, model, or analysis done faster with fewer handoffs.

The best AI outputs come from detailed prompts. But typing long, context-rich prompts is slow - so most people don't bother.

Wispr Flow turns your voice into clean, ready-to-paste text. Speak naturally into ChatGPT, Claude, Cursor, or any AI tool and get polished output without editing. Describe edge cases, explain context, walk through your thinking - all at the speed you talk.

Millions of people use Flow to give AI tools 10x more context in half the time. 89% of messages sent with zero edits.

Works system-wide on Mac, Windows, iPhone, and now Android (free and unlimited on Android during launch).

Try Wispr Flow free →

Evolving AI: The Pentagon labeled Anthropic a supply chain risk after a dispute over military AI use.

The U.S. Department of Defense informed Anthropic that its systems now carry a supply chain risk label, a designation usually reserved for foreign adversaries. The conflict follows Anthropic’s refusal to allow its models to support domestic surveillance or fully autonomous weapons without human oversight. The ruling could disrupt military programs that rely on Claude through Palantir’s Maven system.

The Pentagon has now turned Anthropic’s refusal on mass surveillance and fully autonomous weapons into a vendor problem, meaning contractors may have to swap models, rewrite approvals, and rethink which lab is safe to build on. At the same time, OpenAI’s new Defense deal leaves room for “all lawful purposes” under human oversight, which shows the market is splitting between labs that want firmer red lines and labs willing to work inside government rules.

Evolving AI: LLM workflows can be complex, opaque, and difficult to secure.

Get the latest ebook from Datadog for practical strategies to monitor, troubleshoot, and protect your LLM applications in production. You’ll get key insights into how to overcome the challenges of deploying LLMs securely and at scale, from debugging multi-step workflows to detecting prompt injection attacks.

Free Download ➜

Evolving AI: Netflix acquires Ben Affleck’s AI filmmaking company InterPositive.

Netflix announced it is acquiring InterPositive, a filmmaking technology company founded by Ben Affleck in 2022. The startup builds AI models that help production teams edit footage from their own productions, correcting issues such as continuity errors, lighting mismatches, or missing shots. The system focuses on post-production assistance rather than synthetic actors. Affleck will join Netflix as a senior adviser as the company continues integrating AI into film and TV workflows.

This deal shows where film AI is actually heading right now: less fake actors, more practical tools that clean up expensive production problems after the shoot. Netflix says InterPositive is built to fix things like continuity gaps, lighting mismatches, background swaps, and missing shots using footage a production already owns, which fits the wider push in entertainment for IP-safe AI that helps crews move faster without handing creative control to a model.

QUICK HITS

🚀 Google's Canvas AI Project-Planning Tool Is Now Available to Everyone in the US.

🔌 Tech firms pledge to pay for AI data centre power costs. But will they?

⚖️ Google faces first lawsuit alleging its AI chatbot encouraged a Florida man to commit suicide.

🎶 Apple Music’s answer to the AI music crisis is to use the honor system.

🧠 Mixing generative AI with physics to create personal items that work in the real world.

 *partner link