---
type: "raw_signal"
id: "2026-in-llms-so-far-760df2"
slug: "2026-in-llms-so-far-760df2"
title: "2026 in LLMs (so far)"
description: "Simon Willison’s annotated September 2026 keynote argues that coding agents crossed a practical reliability threshold when Claude Opus 4.5 and GPT-5.1 were paired with their respective agent harnesses."
retrieval_nugget: "Simon Willison’s annotated September 2026 keynote argues that coding agents crossed a practical reliability threshold when Claude Opus 4.5 and GPT-5.1 were paired with their respective agent harnesses."
observed_at: "2026-09-28"
record_date: "2026-09-28"
date_kind: "observed_at"
why_it_matters: "Public source: https://simonwillison.net/2026/Sep/27/2026-in-llms-so-far"
novelty: "notable"
verification_level: "source-linked"
signal_type: "announcement"
source_platform: "simonwillison.net"
topics: ["claude","coding-agents"]
entities: []
related_patterns: []
source_url: "https://simonwillison.net/2026/Sep/27/2026-in-llms-so-far"
source_urls: ["https://simonwillison.net/2026/Sep/27/2026-in-llms-so-far"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:2026-in-llms-so-far-760df2"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/signals/2026-in-llms-so-far-760df2/","markdown":"https://newruntime.com/signals/2026-in-llms-so-far-760df2.md","json":"https://newruntime.com/signals/2026-in-llms-so-far-760df2.json"}
---

# 2026 in LLMs (so far)

## Retrieval answer

Simon Willison’s annotated September 2026 keynote argues that coding agents crossed a practical reliability threshold when Claude Opus 4.5 and GPT-5.1 were paired with their respective agent harnesses.

Simon Willison’s annotated September 2026 keynote argues that coding agents crossed a practical reliability threshold when Claude Opus 4.5 and GPT-5.1 were paired with their respective agent harnesses. He characterizes the change as a move from frequent mistakes to tools usable day to day.

The same source places that capability shift beside an unresolved deployment constraint: sandboxing and agent security. Willison notes that roughly 40 of 277 conference sessions touched those subjects, while the serious agent-security incident he had predicted had not materialized.

Public source: https://simonwillison.net/2026/Sep/27/2026-in-llms-so-far

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Claude](https://newruntime.com/topics/claude.md): Explore the claude topic hub.
- topic hub: [Coding Agents](https://newruntime.com/topics/coding-agents.md): Explore the coding-agents topic hub.
- collection: [Raw Signals](https://newruntime.com/signals.json): Open every source-linked raw signal.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)