GenAI PMGPT 5.4
A GPT model variant used here for scientific reasoning and agentic chemistry experimentation. The newsletter frames it as a model capable of proposing experimental improvements and driving benchmarked workflows.
Key Highlights
- GPT 5.4 was positioned as a long-context, tool-using model suited for coding, PM reasoning, and autonomous workflows.
- OpenAI expanded the family with GPT-5.4 mini and nano for lower-latency and lower-resource deployment scenarios.
- The model was used inside Codex, Amp Deep mode, and OpenClaw to support planning, code generation, and multi-agent execution.
- In a chemistry workflow with Molecule.one’s Maria, GPT-5.4 proposed experimental improvements that contributed to higher reaction yields.
- For AI PMs, GPT 5.4 is most relevant as a benchmark for workflow orchestration, measurable autonomy, and model-tier product decisions.
GPT 5.4
Overview
GPT 5.4 is an OpenAI model family positioned in the newsletter coverage as a high-capability tool for long-context reasoning, agentic workflows, coding, and scientific experimentation. Across mentions, it appears both as a general-purpose frontier model and as a practical workhorse inside products like ChatGPT, Codex, Amp, and specialized research systems. The model is associated with a 1M-token context window, stronger tool use, improved coding throughput, and variants such as GPT-5.4 mini, nano, and Thinking.For AI Product Managers, GPT 5.4 matters because it shows how a foundation model can move from “chat assistant” to “workflow engine.” In the newsletter, it is used to generate production code, power multi-agent operating setups, support PM-style reasoning and architecture tasks, and even help design chemistry experiments that improved reaction yields in a lab setting. That breadth makes GPT 5.4 relevant not just as a model choice, but as a signal of where product design is going: longer horizons, deeper planning, better tool orchestration, and more autonomous execution.
Key Developments
- 2026-03-06: OpenAI introduced GPT-5.4, highlighting it as a new model release with improved capabilities and broader application potential versus prior versions.
- 2026-03-07: Newsletter coverage emphasized GPT 5.4’s reported 1M-token context window, auto-compaction, smarter tool-calling for on-demand skill loading, and up to 1.5× faster coding performance. GPT-5.4 Thinking was also cited in one-shot creation workflows using Codex.
- 2026-03-08: Dharmesh Shah described GPT 5.4 as especially strong for PM reasoning, long-range execution, and back-end architecture work.
- 2026-03-09: GPT-5.4 was noted as officially shipped by OpenAI, reinforcing its status as a major product release in the model landscape.
- 2026-03-18: OpenAI announced GPT-5.4 mini and GPT-5.4 nano, smaller family variants aimed at lower-latency, lower-resource deployment while retaining advanced capabilities. Coverage also noted availability in ChatGPT, Codex, and the API, optimized for coding, computer use, multimodal understanding, and subagents.
- 2026-03-27: Amp added GPT-5.4 to its new Deep agent mode, tuning it for more Codex-like behavior with longer-form and more code-focused reasoning.
- 2026-03-30: Claire Vo used GPT-5.4 alongside Opus-4.6 and Sonnet-4.6 in OpenClaw, where role-based agents automated outreach, scheduling, and operating tasks via Telegram bots across multiple macOS machines.
- 2026-04-03: Simon Willison referenced GPT-5.4 as part of modern agentic engineering patterns, including red/green TDD, thin project templates, and reusable GitHub-driven workflows for higher software productivity.
- 2026-04-06: The Codex team showed how they use GPT 5.4, Codex Spark, and the Codex app’s plan mode with an open-source Rust harness to one-shot generate and iterate code features, including live workflows reaching 1,200 edits per second in fast mode.
- 2026-06-18: GPT-5.4, paired with Molecule.one’s Maria, helped drive a near-autonomous medicinal chemistry workflow. It proposed TEMPO as an additive for Chan–Lam coupling of primary sulfonamides, generated experimental grids for 10,080 reactions, and contributed to measurable yield gains across two optimization cycles.
Relevance to AI PMs
1. Model selection for agentic products: GPT 5.4’s positioning across coding, tool use, long-context reasoning, and multi-agent setups makes it a useful benchmark when deciding between a flagship model and lighter variants like mini or nano. PMs can map product requirements—latency, context length, autonomy level, and cost—to the right model tier.2. Designing workflow-first experiences: The newsletter examples show GPT 5.4 succeeding when embedded in systems such as Codex, Amp Deep mode, OpenClaw, and Maria Lab. For PMs, the lesson is that value often comes less from raw chat quality and more from orchestration features like plan mode, tool routing, subagents, memory management, and execution loops.
3. Operationalizing measurable autonomy: GPT 5.4 appears in benchmarkable workflows with concrete outcomes: coding throughput, automated outreach, and reaction-yield improvement. AI PMs can use this as a template for product metrics—task completion rate, quality uplift, cycle-time reduction, and human-confirmed success—instead of relying only on subjective model evaluations.
Related
- OpenAI: Creator of GPT-5.4 and the broader model family, including mini and nano variants.
- ChatGPT / chatgpt: One of the product surfaces where GPT-5.4 was reported as available, signaling end-user and prosumer access.
- Codex / openais-codex / codex-spark: Closely linked developer products and model experiences where GPT 5.4 is used for planning, code generation, and high-speed iterative editing.
- Amp: Integrated GPT-5.4 into Deep agent mode for more deliberate coding workflows.
- Claude Code, Claude Opus 4.5, Opus-4.6, Sonnet-4.6: Competing or complementary coding/agent models frequently mentioned alongside GPT-5.4 in agentic engineering and multi-model stacks.
- Simon Willison, Dharmesh Shah, Claire Vo: Influential operators and builders whose usage examples framed GPT 5.4’s practical strengths in software, PM work, and operations.
- OpenClaw: An agent operating environment where GPT-5.4 was used as part of role-based automation.
- Molecule.one / Maria / LifesciBench: Scientific and experimental contexts connecting GPT-5.4 to chemistry reasoning, lab automation, and benchmarked research workflows.
- GPT-5.1 and GPT-5.3: Adjacent model generations used as comparison points for coding reliability and release sequencing.
- Lovable, Cursor, ChatPRD: Related AI product-building tools mentioned in the same ecosystem of PM, prototyping, and coding workflows.
Newsletter Mentions (10)
“A near-autonomous AI chemist improves a challenging reaction in medicinal chemistry - GPT‑5.4 paired with Molecule.one’s Maria proposed using TEMPO as an additive to improve Chan–Lam coupling of primary sulfonamides and generated experimental grids that were run (10,080 reactions) in Maria Lab.”
#1 📝 OpenAI News A near-autonomous AI chemist improves a challenging reaction in medicinal chemistry - GPT‑5.4 paired with Molecule.one’s Maria proposed using TEMPO as an additive to improve Chan–Lam coupling of primary sulfonamides and generated experimental grids that were run (10,080 reactions) in Maria Lab. Across two cycles the mean yield rose from 16.6% to 25.2%, yields improved for 88% of boronic acids and 83% of sulfonamides tested, the share of reactions >30% yield increased from 15.6% to 37.5%, and human bench repeats confirmed higher yields for 11 of 14 substrate pairs (most showing >2× increases).
“Alex and Romain demonstrate how the Codex team uses GPT 5.4, the Codex Spark model, and the Codex app’s plan mode—backed by an open-source Rust harness—to one-shot generate and iterate code features like a NASA Artemis iOS screen and a 2D game at up to 1,200 edits per second.”
#2 ▶️ How OpenAI's Codex Team Builds with Codex (43 Min) | Alex & Romain Peter Yang Alex and Romain demonstrate how the Codex team uses GPT 5.4, the Codex Spark model, and the Codex app’s plan mode—backed by an open-source Rust harness—to one-shot generate and iterate code features like a NASA Artemis iOS screen and a 2D game at up to 1,200 edits per second. The Codex team writes specs in under 10 bullet points when implementing new features, relying on Codex to handle most of the coding work. In “fast mode” with Codex Spark, live edits to a 2D game rendered at an average throughput of 1,200 code changes per second. The Codex app, VS Code extension, and CLI all communicate with the same open-source Rust-based harness, allowing multiple parallel agent tasks independent of a single workspace folder.
“Simon Willison details agentic engineering patterns—using coding agents like Claude Code and GPT-5.4 for red/green TDD, thin project templates, and public GitHub hoarding—to boost software productivity and reliability.”
▶️ Why AI came for coders first, automation timelines, and how we’re inside the AI inflection Lennys Podcast Simon Willison details agentic engineering patterns—using coding agents like Claude Code and GPT-5.4 for red/green TDD, thin project templates, and public GitHub hoarding—to boost software productivity and reliability. GPT-5.1 and Claude Opus 4.5 released in November 2025 advanced coding agents from “mostly working” to “almost always following instructions,” enabling engineers to churn out up to 10,000 lines of code per day. Invoking the prompt “red/green TDD” directs agents to write tests first, run them to confirm failure, implement the code, then rerun tests to confirm success. Willison’s GitHub repositories include simonw/tools with 193 HTML/JavaScript client-side utilities and simonw/ressearch with 75 AI-driven research projects to hoard reusable code experiments.
“Claire Vo installed OpenClaw via a one-line Homebrew script on separate macOS machines (three Mac minis and one MacBook Air), configured nine role-based agents (Polly, Finn, Sam, etc.) using Opus-4.6, Sonnet-4.6 and GPT-5.4 models, and linked them to Telegram bots for automating her business outreach and family scheduling.”
#1 ▶️ How OpenClaw’s AI agents run this founder’s business, family and life | Claire Vo Lennys Podcast Claire Vo installed OpenClaw via a one-line Homebrew script on separate macOS machines (three Mac minis and one MacBook Air), configured nine role-based agents (Polly, Finn, Sam, etc.) using Opus-4.6, Sonnet-4.6 and GPT-5.4 models, and linked them to Telegram bots for automating her business outreach and family scheduling. She ran “brew install openclaw” in iTerm, chose personal use, selected Opus-4.6, Sonnet-4.6 and GPT-5.4, then registered each agent as a Telegram bot via BotFather. Agent “Sam” performs a daily sweep of her CRM for product-led growth signups, enriches leads with Exa People Search, drafts and sends outreach emails via Telegram, replacing a human assistant who worked 10 hours/week. She enabled macOS Screen Sharing and Remote Login on her Mac minis to SSH into and view the agent GUIs from her laptop over Wi-Fi, removing the need for dedicated monitors, keyboards or mice.
“Amp has placed GPT-5.4 into its new Deep agent mode, tuning the model to behave more like Codex for longer-form, more code-focused reasoning.”
#6 📝 Ampcode Chronicle GPT‐5.4 in Deep - Amp has placed GPT-5.4 into its new Deep agent mode, tuning the model to behave more like Codex for longer-form, more code-focused reasoning. The update emphasizes deeper planning and agentic behavior for coding tasks.
“OpenAI introduces GPT-5.4 mini and nano - OpenAI announces GPT-5.4 mini and nano, smaller variants of the GPT-5.4 family designed for more efficient deployment while retaining advanced capabilities.”
#1 📝 OpenAI News Introducing GPT-5.4 mini and nano - OpenAI announces GPT-5.4 mini and nano, smaller variants of the GPT-5.4 family designed for more efficient deployment while retaining advanced capabilities. The release targets use cases needing lower latency and resource usage. Also covered by: @Simon Willison #2 𝕏 OpenAI released GPT-5.4 mini today in ChatGPT, Codex and the API—optimized for coding, computer use, multimodal understanding and subagents.
“OpenAI shipped GPT-5.4, Anthropic released a free AI course library, and a browser-based spy-satellite simulator debuted.”
GenAI PM Daily March 09, 2026 GenAI PM Daily 🎧 Listen to this brief 3 min listen Today's top 9 insights for PM Builders, ranked by relevance from X, YouTube, and LinkedIn. OpenAI Ships GPT-5.4 Model #1 𝕏 There's An AI For That : OpenAI shipped GPT-5.4, Anthropic released a free AI course library, and a browser-based spy-satellite simulator debuted. A rogue AI agent went off-script, AI fakes flooded Iran war coverage, and Claude Cowork got a full walkthrough.
“in Dharmesh Shah Dharmesh Shah finds GPT 5.4 excels as both PM (reasoning, long-range execution) and back-end architect (deep thinking, precise execution).”
in Dharmesh Shah Dharmesh Shah finds GPT 5.4 excels as both PM (reasoning, long-range execution) and back-end architect (deep thinking, precise execution). He sees Lovable as the go-to UX designer for polished prototypes and Opus 4.
“OpenAI Releases GPT-5.4 with 1M Token Context #1 in Dharmesh Shah announces OpenAI’s GPT 5.4 launch, featuring a 1 million-token context window with auto-compaction, smarter tool-calling for on-demand skill loading, and up to 1.5× faster coding performance—enabling new HubSpot data-dictionary use cases.”
GenAI PM Daily March 07, 2026 GenAI PM Daily 🎧 Listen to this brief 3 min listen Today's top 25 insights for PM Builders, ranked by relevance from LinkedIn, YouTube, X, and Blogs. OpenAI Releases GPT-5.4 with 1M Token Context #1 in Dharmesh Shah announces OpenAI’s GPT 5.4 launch, featuring a 1 million-token context window with auto-compaction, smarter tool-calling for on-demand skill loading, and up to 1.5× faster coding performance—enabling new HubSpot data-dictionary use cases. Also covered by: @LlamaIndex 🦙 #2 ▶️ What the New ChatGPT 5.4 Means for the World AI Explained GPT-5.4 Thinking, released 48 hours after GPT-5.3 Instant, demonstrated one-shot creation of an animated league table for Stockport County FC using OpenAI’s Codex on Windows and Mac.
“OpenAI Introduces GPT-5.4 Model #1 📝 OpenAI News Introducing GPT-5.4 - Announcement of GPT-5.4 as a new product release, highlighting improvements and new capabilities over prior models. The post introduces features and potential applications of GPT-5.4.”
GenAI PM Daily March 06, 2026 GenAI PM Daily 🎧 Listen to this brief 3 min listen Today's top 25 insights for PM Builders, ranked by relevance from Blogs, X, LinkedIn, and YouTube. OpenAI Introduces GPT-5.4 Model #1 📝 OpenAI News Introducing GPT-5.4 - Announcement of GPT-5.4 as a new product release, highlighting improvements and new capabilities over prior models. The post introduces features and potential applications of GPT-5.4. Also covered by: @There's An AI For That , @Kevin Weil 🇺🇸 #2 𝕏 claire vo 🖤 GPT-5.4 just went live in @chatprd with a 1M-token context window, more human-like dialogue than 5.2/5.3, and chef’s-kiss tool use for deep investigations. She flags it still defaults to bullet points, needs front-end/UX polish, and has latency/stability TBD.
Related
An AI coding assistant environment used for running evaluation skills and agentic workflows. In this issue it is mentioned as a runtime for ai-evals-course material and as an agent in an OpenRouter-like system.
An AI company building frontier models, ChatGPT, and custom inference hardware. Here it is discussed for Jalapeño and ChatGPT Business Premium Seats.
An AI coding tool referenced as providing data used to evaluate Grok 4.6. It is also named later as a target environment for running AI eval skills.
An AI coding agent or environment mentioned as a place to run AI eval skills. It is also listed as one of the agents that can be compared in a shared environment.
A prominent AI blogger and commentator referenced in connection with an article on token reselling and fraud. He is cited as the source of the newsletter item discussing the marketplace and API-key abuse.
A standardized agent test suite referenced for model evaluation. The newsletter cites success rates on OpenClaw as part of the Nemotron benchmark result.
OpenAI’s conversational AI product used by the design team to prototype ideas and test interface decisions. Here it is also part of a rapid experimentation workflow.
An operator or product thinker who raised concerns about data indexing, connector visibility, prompt injection, and evaluation quality. Her comment focuses on trust, deletion, and user-empathetic system design.
Co-founder associated here with advocating an 'open brain' approach to machine- and human-readable organizational information. Important for PMs thinking about internal systems, APIs, and organizational memory.
An agent platform whose agents can schedule wake-ups, retain context, and trigger workflows. Useful for PMs exploring persistent, scheduled AI automation tied into collaboration tools.
A Claude model version praised for personality and writing style. The newsletter contrasts it with Opus 5 as more concise and friend-like.
An AI-first product management tool or startup referenced by Claire Vo. The newsletter uses it in a discussion of shipping an AI-first version of an app without traditional PM tooling.
A no-code AI app builder referenced here as the platform used to build a production-grade SaaS product. For PMs, it illustrates how agentic coding is changing build-vs-buy and software creation economics.
A Claude model used in the newsletter's example to run Python code and analyze a floor plan. It is discussed as part of an agentic workflow inside Claude Cowork.
A test-driven development pattern adapted for coding agents. It emphasizes an iterative failure/success loop that can make agentic coding more reliable.
Stay updated on GPT 5.4
Get curated AI PM insights delivered daily — covering this and 1,000+ other sources.
Subscribe Free