Michelle Pokrass and Josh McGrath from OpenAI dive into the exciting updates of GPT 4.1. They discuss its enhanced coding capabilities and instruction-following features, making it a developer's new best friend. The conversation touches on the innovative Nano model designed for low latency, and they share the fun of naming projects. With insights into pricing and user interaction, they emphasize the significance of community feedback in evolving AI technology. Plus, get ready for the intriguing benefits of multimodal tasks and cutting-edge reasoning enhancements!
41:52
forum Ask episode
web_stories AI Snips
view_agenda Chapters
auto_awesome Transcript
info_circle Episode notes
00:00 / 00:00
Codenames and Developer Feedback
OpenAI used codenames like "Quasar" and "Optimus" for GPT 4.1 during testing through OpenRouter.
This allowed for gathering valuable developer feedback while keeping the model's identity somewhat disguised.
00:00 / 00:00
GPT 4.1 Positioning
GPT 4.1 is a significant improvement over GPT 4.0, but smaller and cheaper than GPT 4.5.
It doesn't beat 4.5 on all evaluations, but most developers can replace 4.5 usage with 4.1.
00:00 / 00:00
Challenges of Long Context
Reaching a 1 million context window was challenging, requiring new evaluations for long context reasoning like MRCR and GraphWalks.
Simple needle-in-a-haystack tasks proved easier for models than complex reasoning across multiple documents.
Get the Snipd Podcast app to discover more snips from this episode
We’ll keep this brief because we’re on a tight turnaround: GPT 4.1, previously known as the Quasar and Optimus models, is now live as the natural update for 4o/4o-mini (and the research preview of GPT 4.5). Though it is a general purpose model family, the headline features are:
Coding abilities (o1-level SWEBench and SWELancer, but ok Aider)
Instruction Following (with a very notable prompting guide)
Long Context up to 1m tokens (with new MRCR and Graphwalk benchmarks)
Vision (simply o1 level)
Cheaper Pricing (cheaper than 4o, greatly improved prompt caching savings)
We caught up with returning guest Michelle Pokrass and Josh McGrath to get more detail on each!
Chapters
00:00:00 Introduction and Guest Welcome
00:00:57 GPC 4.1 Launch Overview
00:01:54 Developer Feedback and Model Names
00:02:53 Model Naming and Starry Themes
00:03:49 Confusion Over GPC 4.1 vs 4.5
00:04:47 Distillation and Model Improvements
00:05:45 Omnimodel Architecture and Future Plans
00:06:43 Core Capabilities of GPC 4.1
00:07:40 Training Techniques and Long Context
00:08:37 Challenges in Long Context Reasoning
00:09:34 Context Utilization in Models
00:10:31 Graph Walks and Model Evaluation
00:11:31 Real Life Applications of Graph Tasks
00:12:30 Multi-Hop Reasoning Benchmarks
00:13:30 Agentic Workflows and Backtracking
00:14:28 Graph Traversals for Agent Planning
00:15:24 Context Usage in API and Memory Systems
00:16:21 Model Performance in Long Context Tasks
00:17:17 Instruction Following and Real World Data
00:18:12 Challenges in Grading Instructions
00:19:09 Instruction Following Techniques
00:20:09 Prompting Techniques and Model Responses
00:21:05 Agentic Workflows and Model Persistence
00:22:01 Balancing Persistence and User Control
00:22:56 Evaluations on Model Edits and Persistence
Latent Space: The AI Engineer Podcast 
