# Sources & repositories

PROVENANCE
Sources & repositories
Follow each work back to its code, creator and original question.
Repositories & demos
23 entry points
Original licenses apply; check each repository’s LICENSE. Unverified source packages are not available for download.
01
svg-set
simonw/pelican-bicycle
The original repository and early SVG outputs.
Source ↗
Demo ↗
ZIP ↓
02
threejs-game
Licoy/tihuqiche
An island cycling game built with Vue 3 and Three.js.
Source ↗
Demo ↗
ZIP ↓
03
svg-set
scosman/pelicans_riding_bicycles
A curated gallery of model-generated SVGs.
Source ↗
Demo ↗
ZIP ↓
04
eval-harness
haowang02/auto-pelican
An agent that generates and evaluates SVG animations.
Source ↗
ZIP ↓
05
blender
simonw/gpt-6-astra-blender-pelican-bicycle
A cycling pelican built in local Blender with Astra.
Source ↗
ZIP ↓
06
app
simonw/pedalican
A cycling pelican desktop pet for Codex.
Source ↗
ZIP ↓
07
svg-grid
JoJohanse/pelican-bicycle-benchmark
SVG outputs compared across multiple models.
Source ↗
Demo ↗
ZIP ↓
08
agent
youngbrioche/agentic-pelican-on-a-bicycle
Iterative pelican drawing with an agent.
Source ↗
Demo ↗
ZIP ↓
09
html-anim
xiiiblue/pelican-bicycle
HTML animations compared across models.
Source ↗
Demo ↗
ZIP ↓
10
svg-animated
taikongren776/pelican-bicycle
A seaside cycling scene in animated SVG.
Source ↗
Demo ↗
ZIP ↓
11
set
XinhanWang/bicycle-pelican
A collection exploring bicycles and pelicans.
Source ↗
Demo ↗
ZIP ↓
12
svg-animated
Superxian-boop/pelican-bicycle
Cycling expressed in pure SVG animation.
Source ↗
Demo ↗
ZIP ↓
13
webapp
heysong20651901/pelican-bike
An SVG-based pelican web app.
Source ↗
Demo ↗
ZIP ↓
14
svg-set
sebseb7/pelicanTest
A collection of model-generated SVG experiments.
Source ↗
Demo ↗
ZIP ↓
15
threejs-game
claude-opus-5-5.riba2534.cn
A playable coastal 3D scene with waves and cycling.
Source ↗
Demo ↗
16
threejs-game
tihuqiche.com
The online edition of the open-source cycling game.
Source ↗
Demo ↗
17
4kb-demo
demobench.janekm.com
A 4 KB demo with GIF, assembly source and a WASM runtime.
Source ↗
Demo ↗
18
html-anim
CHNAICORP/pelican-ride
A single-file SVG and HTML animation.
Source ↗
Demo ↗
ZIP ↓
19
html-set
vastxie/pelican-gallery-data
Original HTML animations and model runs; 164 extracted files.
Source ↗
Demo ↗
ZIP ↓
20
svg-animated
Vulcan575/pelican-riding-bicycle
A dependency-free teal city bicycle animation.
Source ↗
Demo ↗
ZIP ↓
21
html-grid
RustFisher · 3-model comparison
Astra, Gemini 3.8 Flash and DeepSeek 4.1 Flash compared.
Source ↗
Demo ↗
22
threejs
ohmyopus · 3D city
An archived preview with Python / pybevy source.
Source ↗
Demo ↗
Code ↗
23
html-anim
prab · upstream unverified
Attributed to DeepSeek V4.1 Flash; upstream repository unverified.
Demo ↗
THE PERSON BEHIND THE PROMPT
Simon
Willison
↗
Django co-creator · Creator of Datasette
About Simon ↗
Why does this pelican lead back to Simon?
On October 25, 2024, Simon published SVG outputs from 16 models using the same prompt. He chose pelicans because he liked them and believed bicycle-riding pelican SVGs were then rare in training data.
Generate an SVG of a pelican riding a bicycle
The community extended this visual experiment into animation, 3D and games. This archive follows that evolving body of work; individual samples do not establish overall model capability.
Read the original experiment ↗
Simon’s related writing
144 entries
Search articles
Year
All years
2026
2025
2024
2026-09-22
Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
2026-09-18
The Creative Spirit of Who Framed Roger Rabbit
2026-09-05
Introducing GPT-6 Astra for developers
2026-09-05
TIL Using Blender with coding agents on macOS I've been having fun with Blender in ChatGPT Codex on my Mac recently. Get
2026-09-04
The Pelican comparison grid for Astra is pretty interesting
2026-09-02
Release llm-gemini 0.34 New model gemini-3.8-flash for Gemini 3.8 Flash , with low, medium and high thinking levels. #14
2026-09-01
Claude Fable 5.1 made me a really nice animated pelican
2026-08-29
Introducing Hy4 Preview
2026-08-26
Qwen3.8-Flash-Next
2026-08-16
Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things
2026-08-13
Release llm-gemini 0.33 It's been a while since the last llm-gemini release. This version of the plugin adds support for
2026-08-12
DeepSeek V4 Pro 0813 (on OpenRouter)
2026-08-10
Introducing Muse Glimmer
2026-08-05
Introducing Muse Code and Muse Spark 1.2
2026-07-31
deepseek-ai/DeepSeek-V4-Flash-0731
2026-07-22
Are AI labs pelicanmaxxing?
2026-07-20
Who’s Afraid of Chinese Models?
2026-07-16
Kimi K3, and what we can still learn from the pelican benchmark
2026-07-16
Inkling: Our open-weights model
2026-07-14
simonw/pedalican
2026-07-09
The new GPT-5.6 family: Luna, Terra, Sol
2026-07-09
Introducing Muse Spark 1.1
2026-07-06
tencent/Hy3
2026-06-30
What’s new in Claude Sonnet 5
2026-06-29
Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding
2026-06-17
GLM-5.2 is probably the most powerful text-only open weights LLM
2026-06-10
DiffusionGemma
2026-06-09
Initial impressions of Claude Fable 5
2026-05-28
Claude Opus 4.8: “a modest but tangible improvement”
2026-05-19
Gemini 3.5 Flash: more expensive, but Google plan to use it for everything
2026-05-19
The last six months in LLMs in five minutes
2026-05-04
Granite 4.1 3B SVG Pelican Gallery
2026-04-25
@scottjla on Twitter in reply to my pelican riding a bicycle benchmark: I feel like we need to stack these tests now I c
2026-04-24
DeepSeek V4—almost on the frontier, a fraction of the price
2026-04-23
A pelican for GPT-5.5 via the semi-official Codex backdoor API
2026-04-22
Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model
2026-04-21
scosman/pelicans_riding_bicycles
2026-04-20
Release llm-openrouter 0.6 llm openrouter refresh command for refreshing the list of available models without waiting fo
2026-04-16
Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7
2026-04-08
Meta’s new model is Muse Spark, and meta.ai chat has some interesting tools
2026-04-07
GLM-5.1: Towards Long-Horizon Tasks
2026-04-02
Gemma 4: Byte for byte, the most capable open models
2026-03-17
GPT-5.4 mini and GPT-5.4 nano, which can describe 76,000 photos for $52
2026-03-16
Introducing Mistral Small 4
2026-03-05
Introducing GPT‑5.4
2026-03-03
Gemini 3.1 Flash-Lite
2026-02-19
Gemini 3.1 Pro
2026-02-17
Introducing Claude Sonnet 4.6
2026-02-17
Opus rear bike frame makes no sense structurally, there is no connection from pedals to rear wheel, both legs of Pelican
2026-02-17
Qwen3.5: Towards Native Multimodal Agents
2026-02-12
Introducing GPT‑5.3‑Codex‑Spark
2026-02-12
Gemini 3 Deep Think
2026-02-11
GLM-5: From Vibe Coding to Agentic Engineering
2026-02-05
Two major new model releases today, within about 15 minutes of each other. Anthropic released Opus 4.6 . Here's its peli
2026-01-27
Kimi K2.5: Visual Agentic Intelligence
2025-12-31
2025: The year in LLMs
2025-12-19
Introducing GPT-5.2-Codex
2025-12-17
Gemini 3 Flash
2025-12-11
GPT-5.2
2025-12-09
Devstral 2
2025-12-01
DeepSeek-V3.2
2025-11-25
LLM SVG Generation Benchmark
2025-11-24
Claude Opus 4.5, and why evaluating new LLMs is increasingly difficult
2025-11-22
Olmo 3 is a fully open LLM
2025-11-19
Building more with GPT-5.1-Codex-Max
2025-11-18
Trying out Gemini 3 Pro with audio transcription and a new pelican benchmark
2025-11-13
Introducing GPT-5.1 for developers
2025-11-13
What happens if AI labs train for pelicans riding bicycles?
2025-11-11
Agentic Pelican on a Bicycle
2025-11-09
Pelican on a Bike—Raytracer Edition
2025-11-09
Reverse engineering Codex CLI to get GPT-5-Codex-Mini to draw me a pelican
2025-11-06
Kimi K2 Thinking
2025-10-29
Introducing SWE-1.5: Our Fast Agent Model
2025-10-29
MiniMax M2 & Agent: Ingenious in Simplicity
2025-10-29
Composer: Building a fast frontier model with RL
2025-10-25
Visual Features Across Modalities: SVG and ASCII Art Reveal Cross-Modal Understanding
2025-10-15
Introducing Claude Haiku 4.5
2025-10-06
gpt-image-1-mini
2025-10-06
GPT-5 pro
2025-10-01
Two new models from Chinese AI labs in the past few days. I tried them both out using llm-openrouter : DeepSeek-V3.2-Exp
2025-09-29
Claude Sonnet 4.5 is probably the “best coding model in the world” (at least for now)
2025-09-25
Improved Gemini 2.5 Flash and Flash-Lite
2025-09-23
GPT-5-Codex
2025-09-20
Grok 4 Fast
2025-09-15
GPT‑5-Codex and upgrades to Codex
2025-09-12
Qwen3-Next-80B-A3B
2025-09-06
Kimi-K2-Instruct-0905
2025-08-22
DeepSeek 3.1
2025-08-20
David Ho on BlueSky: A pelican tried to eat my bike
2025-08-14
Introducing Gemma 3 270M: The compact model for hyper-efficient AI
2025-08-10
Qwen3-4B-Thinking: “This is art—pelicans don’t ride bikes!”
2025-08-07
A couple of weeks ago I was invited to OpenAI's headquarters for a "preview event", for which I had to sign both an NDA
2025-08-07
GPT-5: Key characteristics, pricing and model card
2025-08-05
OpenAI’s new open weight (Apache 2) models are really good
2025-08-05
Claude Opus 4.1
2025-08-03
XBai o4
2025-08-01
Deep Think in the Gemini app
2025-07-31
Trying out Qwen3 Coder Flash using LM Studio and Open WebUI and LLM
2025-07-30
Qwen3-30B-A3B-Thinking-2507
2025-07-29
My 2.5 year old laptop can write Space Invaders in JavaScript now, using GLM-4.5 Air and MLX
2025-07-28
GLM-4.5: Reasoning, Coding, and Agentic Abililties
2025-07-25
Qwen3-235B-A22B-Thinking-2507
2025-07-23
Instagram Reel: Veo 3 paid preview
2025-07-22
Qwen3-Coder: Agentic Coding in the World
2025-07-22
Qwen/Qwen3-235B-A22B-Instruct-2507
2025-07-11
moonshotai/Kimi-K2-Instruct
2025-07-10
Grok 4
2025-07-09
Infinite Monkey
2025-06-26
Introducing Gemma 3n: The developer guide
2025-06-20
Mistral-Small 3.2
2025-06-17
Trying out the new Gemini 2.5 model family
2025-06-10
o3-pro
2025-06-10
Magistral — the first reasoning model by Mistral AI
2025-06-06
The last six months in LLMs, illustrated by pelicans on bicycles
2025-05-22
Live blog: Claude 4 launch at Code with Claude
2025-05-20
Gemini 2.5: Our most intelligent models are getting even better
2025-05-20
Tucked into today's Google I/O keynote, a blink-and-you'll miss it moment : The pelican in the keynote was created by Al
2025-05-17
In addition to my workshop the other day I'm also participating in the poster session at PyCon US this year. This means
2025-05-06
Gemini 2.5 Pro Preview: even better coding performance
2025-04-29
Qwen 3 offers a case study in how to effectively release a model
2025-04-17
Start building with Gemini 2.5 Flash
2025-04-14
GPT-4.1: Three new million token input models from OpenAI, including their cheapest model yet
2025-03-25
Putting Gemini 2.5 Pro through its paces
2025-03-24
deepseek-ai/DeepSeek-V3-0324
2025-03-19
OpenAI platform: o1-pro
2025-03-16
mlx-community/OLMo-2-0325-32B-Instruct-4bit
2025-03-12
Notes on Google’s Gemma 3
2025-02-27
Initial impressions of GPT-4.5
2025-02-24
Claude 3.7 Sonnet and Claude Code
2025-02-18
Andrej Karpathy’s initial impressions of Grok 3
2025-02-05
Gemini 2.0 is now available to everyone
2025-01-20
DeepSeek-R1 and exploring DeepSeek-R1-Distill-Llama-8B
2025-01-08
microsoft/phi-4
2024-12-19
Gemini 2.0 Flash “Thinking mode”
2024-12-16
Veo 2
2024-12-13
<model-viewer> Web Component by Google
2024-12-09
Sora
2024-12-09
I can now run a GPT-4 class model on my laptop
2024-12-06
New Gemini model: gemini-exp-1206
2024-11-27
QwQ: Reflect Deeply on the Boundaries of the Unknown
2024-11-22
Say hello to gemini-exp-1121
2024-11-15
Recraft V3
2024-11-12
Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac
2024-10-25
Pelicans on a bicycle
No matching articles.
Scroll within the list to browse. These reading references are not counted as artworks.

## Links
- Source ↗: https://github.com/simonw/pelican-bicycle
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/simon-origin/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/simon-origin.zip
- Source ↗: https://github.com/Licoy/tihuqiche
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/tihuqiche/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/tihuqiche.zip
- Source ↗: https://github.com/scosman/pelicans_riding_bicycles
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/scosman/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/scosman.zip
- Source ↗: https://github.com/haowang02/auto-pelican
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/haowang-auto.zip
- Source ↗: https://github.com/simonw/gpt-6-astra-blender-pelican-bicycle
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/blender-astra.zip
- Source ↗: https://github.com/simonw/pedalican
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/pedalican.zip
- Source ↗: https://github.com/JoJohanse/pelican-bicycle-benchmark
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/jojohanse/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/jojohanse.zip
- Source ↗: https://github.com/youngbrioche/agentic-pelican-on-a-bicycle
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/youngbrioche/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/youngbrioche.zip
- Source ↗: https://github.com/xiiiblue/pelican-bicycle
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/xiiiblue/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/xiiiblue.zip
- Source ↗: https://github.com/taikongren776/pelican-bicycle
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/taikongren/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/taikongren.zip
- Source ↗: https://github.com/XinhanWang/bicycle-pelican
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/xinhan/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/xinhan.zip
- Source ↗: https://github.com/Superxian-boop/pelican-bicycle
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/superxian/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/superxian.zip
- Source ↗: https://github.com/heysong20651901/pelican-bike
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/heysong/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/heysong.zip
- Source ↗: https://github.com/sebseb7/pelicanTest
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/sebseb7/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/sebseb7.zip
- Source ↗: https://claude-opus-5-5.riba2534.cn/
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/riba-opus-coast.html
- Source ↗: https://tihuqiche.com/
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/tihuqiche/index.html
- Source ↗: https://demobench.janekm.com/
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/demobench/index.html
- Source ↗: https://github.com/CHNAICORP/pelican-ride
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/CHNAICORP-pelican-ride/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/CHNAICORP-pelican-ride.zip
- Source ↗: https://github.com/vastxie/pelican-gallery-data
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/vastxie-gallery/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/vastxie-gallery.zip
- Source ↗: https://github.com/Vulcan575/pelican-riding-bicycle
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/Vulcan575/index.html
- ZIP ↓: https://pelicanmap.aveniqa.com/downloads/Vulcan575.zip
- Source ↗: https://en.rustfisher.com/ai/pelican-bike-demo202609-gpt6astra-gemini3-8flash-deepseek4-1flash/
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/sites/rustfisher-compare.html
- Source ↗: https://ohmyopus.com/en/c/pelican-threejs
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/sites/ohmyopus.html
- Code ↗: https://pelicanmap-demos.aveniqa.com/demos/sites/ohmyopus-pelican.py
- Demo ↗: https://pelicanmap-demos.aveniqa.com/demos/staypzy/index.html
- About Simon ↗: https://simonwillison.net/about/
- Read the original experiment ↗: https://simonwillison.net/2024/Oct/25/pelicans-on-a-bicycle/
- Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war: https://simonwillison.net/2026/Sep/22/opus-and-sol-and-luna/
- The Creative Spirit of Who Framed Roger Rabbit: https://simonwillison.net/2026/Sep/18/the-creative-spirit-of-who-framed-roger-rabbit/
- Introducing GPT-6 Astra for developers: https://simonwillison.net/2026/Sep/5/introducing-gpt-6-astra-for-developers/
- TIL Using Blender with coding agents on macOS I've been having fun with Blender in ChatGPT Codex on my Mac recently. Get: https://simonwillison.net/2026/Sep/5/blender-coding-agents-macos/
- The Pelican comparison grid for Astra is pretty interesting: https://simonwillison.net/2026/Sep/4/astra-pelicans/
- Release llm-gemini 0.34 New model gemini-3.8-flash for Gemini 3.8 Flash , with low, medium and high thinking levels. #14: https://simonwillison.net/2026/Sep/2/llm-gemini/
- Claude Fable 5.1 made me a really nice animated pelican: https://simonwillison.net/2026/Sep/1/claude-fable-5-1/
- Introducing Hy4 Preview: https://simonwillison.net/2026/Aug/29/hy4/
- Qwen3.8-Flash-Next: https://simonwillison.net/2026/Aug/26/qwen38-flash-next/
- Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things: https://simonwillison.net/2026/Aug/16/qwen-38-27b/
- Release llm-gemini 0.33 It's been a while since the last llm-gemini release. This version of the plugin adds support for: https://simonwillison.net/2026/Aug/13/llm-gemini/
- DeepSeek V4 Pro 0813 (on OpenRouter): https://simonwillison.net/2026/Aug/12/deepseek-v4-pro-0813/
- Introducing Muse Glimmer: https://simonwillison.net/2026/Aug/10/introducing-muse-glimmer/
- Introducing Muse Code and Muse Spark 1.2: https://simonwillison.net/2026/Aug/5/muse-code-and-muse-spark-12/
- deepseek-ai/DeepSeek-V4-Flash-0731: https://simonwillison.net/2026/Jul/31/deepseek-v4-flash-0731/
- Are AI labs pelicanmaxxing?: https://simonwillison.net/2026/Jul/22/are-ai-labs-pelicanmaxxing/
- Who’s Afraid of Chinese Models?: https://simonwillison.net/2026/Jul/20/afraid-of-chinese-models/
- Kimi K3, and what we can still learn from the pelican benchmark: https://simonwillison.net/2026/Jul/16/kimi-k3/
- Inkling: Our open-weights model: https://simonwillison.net/2026/Jul/16/inkling/
- simonw/pedalican: https://simonwillison.net/2026/Jul/14/pedalican/
- The new GPT-5.6 family: Luna, Terra, Sol: https://simonwillison.net/2026/Jul/9/gpt-5-6/
- Introducing Muse Spark 1.1: https://simonwillison.net/2026/Jul/9/muse-spark-1-1/
- tencent/Hy3: https://simonwillison.net/2026/Jul/6/hy3/
- What’s new in Claude Sonnet 5: https://simonwillison.net/2026/Jun/30/claude-sonnet-5/
- Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding: https://simonwillison.net/2026/Jun/29/ornith/
- GLM-5.2 is probably the most powerful text-only open weights LLM: https://simonwillison.net/2026/Jun/17/glm-52/
- DiffusionGemma: https://simonwillison.net/2026/Jun/10/diffusiongemma/
- Initial impressions of Claude Fable 5: https://simonwillison.net/2026/Jun/9/claude-fable-5/
- Claude Opus 4.8: “a modest but tangible improvement”: https://simonwillison.net/2026/May/28/claude-opus-4-8/
- Gemini 3.5 Flash: more expensive, but Google plan to use it for everything: https://simonwillison.net/2026/May/19/gemini-35-flash/
- The last six months in LLMs in five minutes: https://simonwillison.net/2026/May/19/5-minute-llms/
- Granite 4.1 3B SVG Pelican Gallery: https://simonwillison.net/2026/May/4/granite-41-3b-svg-pelican-gallery/
- @scottjla on Twitter in reply to my pelican riding a bicycle benchmark: I feel like we need to stack these tests now I c: https://simonwillison.net/2026/Apr/25/why-are-you-like-this/
- DeepSeek V4—almost on the frontier, a fraction of the price: https://simonwillison.net/2026/Apr/24/deepseek-v4/
- A pelican for GPT-5.5 via the semi-official Codex backdoor API: https://simonwillison.net/2026/Apr/23/gpt-5-5/
- Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model: https://simonwillison.net/2026/Apr/22/qwen36-27b/
- scosman/pelicans_riding_bicycles: https://simonwillison.net/2026/Apr/21/scosman/
- Release llm-openrouter 0.6 llm openrouter refresh command for refreshing the list of available models without waiting fo: https://simonwillison.net/2026/Apr/20/llm-openrouter/
- Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7: https://simonwillison.net/2026/Apr/16/qwen-beats-opus/
- Meta’s new model is Muse Spark, and meta.ai chat has some interesting tools: https://simonwillison.net/2026/Apr/8/muse-spark/
- GLM-5.1: Towards Long-Horizon Tasks: https://simonwillison.net/2026/Apr/7/glm-51/
- Gemma 4: Byte for byte, the most capable open models: https://simonwillison.net/2026/Apr/2/gemma-4/
- GPT-5.4 mini and GPT-5.4 nano, which can describe 76,000 photos for $52: https://simonwillison.net/2026/Mar/17/mini-and-nano/
- Introducing Mistral Small 4: https://simonwillison.net/2026/Mar/16/mistral-small-4/
- Introducing GPT‑5.4: https://simonwillison.net/2026/Mar/5/introducing-gpt54/
- Gemini 3.1 Flash-Lite: https://simonwillison.net/2026/Mar/3/gemini-31-flash-lite/
- Gemini 3.1 Pro: https://simonwillison.net/2026/Feb/19/gemini-31-pro/
- Introducing Claude Sonnet 4.6: https://simonwillison.net/2026/Feb/17/claude-sonnet-46/
- Opus rear bike frame makes no sense structurally, there is no connection from pedals to rear wheel, both legs of Pelican: https://simonwillison.net/2026/Feb/17/elon-musk/
- Qwen3.5: Towards Native Multimodal Agents: https://simonwillison.net/2026/Feb/17/qwen35/
- Introducing GPT‑5.3‑Codex‑Spark: https://simonwillison.net/2026/Feb/12/codex-spark/
- Gemini 3 Deep Think: https://simonwillison.net/2026/Feb/12/gemini-3-deep-think/
- GLM-5: From Vibe Coding to Agentic Engineering: https://simonwillison.net/2026/Feb/11/glm-5/
- Two major new model releases today, within about 15 minutes of each other. Anthropic released Opus 4.6 . Here's its peli: https://simonwillison.net/2026/Feb/5/two-new-models/
- Kimi K2.5: Visual Agentic Intelligence: https://simonwillison.net/2026/Jan/27/kimi-k25/
- 2025: The year in LLMs: https://simonwillison.net/2025/Dec/31/the-year-in-llms/
- Introducing GPT-5.2-Codex: https://simonwillison.net/2025/Dec/19/introducing-gpt-52-codex/
- Gemini 3 Flash: https://simonwillison.net/2025/Dec/17/gemini-3-flash/
- GPT-5.2: https://simonwillison.net/2025/Dec/11/gpt-52/
- Devstral 2: https://simonwillison.net/2025/Dec/9/devstral-2/
- DeepSeek-V3.2: https://simonwillison.net/2025/Dec/1/deepseek-v32/
- LLM SVG Generation Benchmark: https://simonwillison.net/2025/Nov/25/llm-svg-generation-benchmark/
- Claude Opus 4.5, and why evaluating new LLMs is increasingly difficult: https://simonwillison.net/2025/Nov/24/claude-opus/
- Olmo 3 is a fully open LLM: https://simonwillison.net/2025/Nov/22/olmo-3/
- Building more with GPT-5.1-Codex-Max: https://simonwillison.net/2025/Nov/19/gpt-51-codex-max/
- Trying out Gemini 3 Pro with audio transcription and a new pelican benchmark: https://simonwillison.net/2025/Nov/18/gemini-3/
- Introducing GPT-5.1 for developers: https://simonwillison.net/2025/Nov/13/gpt-51/
- What happens if AI labs train for pelicans riding bicycles?: https://simonwillison.net/2025/Nov/13/training-for-pelicans-riding-bicycles/
- Agentic Pelican on a Bicycle: https://simonwillison.net/2025/Nov/11/agentic-pelican-on-a-bicycle/
- Pelican on a Bike—Raytracer Edition: https://simonwillison.net/2025/Nov/9/pelican-on-a-bike-raytracer-edition/
- Reverse engineering Codex CLI to get GPT-5-Codex-Mini to draw me a pelican: https://simonwillison.net/2025/Nov/9/gpt-5-codex-mini/
- Kimi K2 Thinking: https://simonwillison.net/2025/Nov/6/kimi-k2-thinking/
- Introducing SWE-1.5: Our Fast Agent Model: https://simonwillison.net/2025/Oct/29/swe-15/
- MiniMax M2 & Agent: Ingenious in Simplicity: https://simonwillison.net/2025/Oct/29/minimax-m2/
- Composer: Building a fast frontier model with RL: https://simonwillison.net/2025/Oct/29/cursor-composer/
- Visual Features Across Modalities: SVG and ASCII Art Reveal Cross-Modal Understanding: https://simonwillison.net/2025/Oct/25/visual-features-across-modalities/
- Introducing Claude Haiku 4.5: https://simonwillison.net/2025/Oct/15/claude-haiku-45/
- gpt-image-1-mini: https://simonwillison.net/2025/Oct/6/gpt-image-1-mini/
- GPT-5 pro: https://simonwillison.net/2025/Oct/6/gpt-5-pro/
- Two new models from Chinese AI labs in the past few days. I tried them both out using llm-openrouter : DeepSeek-V3.2-Exp: https://simonwillison.net/2025/Oct/1/two-pelicans/
- Claude Sonnet 4.5 is probably the “best coding model in the world” (at least for now): https://simonwillison.net/2025/Sep/29/claude-sonnet-4-5/
- Improved Gemini 2.5 Flash and Flash-Lite: https://simonwillison.net/2025/Sep/25/improved-gemini-25-flash-and-flash-lite/
- GPT-5-Codex: https://simonwillison.net/2025/Sep/23/gpt-5-codex/
- Grok 4 Fast: https://simonwillison.net/2025/Sep/20/grok-4-fast/
- GPT‑5-Codex and upgrades to Codex: https://simonwillison.net/2025/Sep/15/gpt-5-codex/
- Qwen3-Next-80B-A3B: https://simonwillison.net/2025/Sep/12/qwen3-next/
- Kimi-K2-Instruct-0905: https://simonwillison.net/2025/Sep/6/kimi-k2-instruct-0905/
- DeepSeek 3.1: https://simonwillison.net/2025/Aug/22/deepseek-31/
- David Ho on BlueSky: A pelican tried to eat my bike: https://simonwillison.net/2025/Aug/20/a-pelican-tried-to-eat-my-bike/
- Introducing Gemma 3 270M: The compact model for hyper-efficient AI: https://simonwillison.net/2025/Aug/14/gemma-3-270m/
- Qwen3-4B-Thinking: “This is art—pelicans don’t ride bikes!”: https://simonwillison.net/2025/Aug/10/qwen3-4b/
- A couple of weeks ago I was invited to OpenAI's headquarters for a "preview event", for which I had to sign both an NDA: https://simonwillison.net/2025/Aug/7/previewing-gpt-5/
- GPT-5: Key characteristics, pricing and model card: https://simonwillison.net/2025/Aug/7/gpt-5/
- OpenAI’s new open weight (Apache 2) models are really good: https://simonwillison.net/2025/Aug/5/gpt-oss/
- Claude Opus 4.1: https://simonwillison.net/2025/Aug/5/claude-opus-41/
- XBai o4: https://simonwillison.net/2025/Aug/3/xbai-o4/
- Deep Think in the Gemini app: https://simonwillison.net/2025/Aug/1/deep-think-in-the-gemini-app/
- Trying out Qwen3 Coder Flash using LM Studio and Open WebUI and LLM: https://simonwillison.net/2025/Jul/31/qwen3-coder-flash/
- Qwen3-30B-A3B-Thinking-2507: https://simonwillison.net/2025/Jul/30/qwen3-30b-a3b-thinking-2507/
- My 2.5 year old laptop can write Space Invaders in JavaScript now, using GLM-4.5 Air and MLX: https://simonwillison.net/2025/Jul/29/space-invaders/
- GLM-4.5: Reasoning, Coding, and Agentic Abililties: https://simonwillison.net/2025/Jul/28/glm-45/
- Qwen3-235B-A22B-Thinking-2507: https://simonwillison.net/2025/Jul/25/qwen3-235b-a22b-thinking-2507/
- Instagram Reel: Veo 3 paid preview: https://simonwillison.net/2025/Jul/23/instagram-reel-veo-3-paid-preview/
- Qwen3-Coder: Agentic Coding in the World: https://simonwillison.net/2025/Jul/22/qwen3-coder/
- Qwen/Qwen3-235B-A22B-Instruct-2507: https://simonwillison.net/2025/Jul/22/qwen3-235b-a22b-instruct-2507/
- moonshotai/Kimi-K2-Instruct: https://simonwillison.net/2025/Jul/11/kimi-k2/
- Grok 4: https://simonwillison.net/2025/Jul/10/grok-4/
- Infinite Monkey: https://simonwillison.net/2025/Jul/9/infinite-monkey/
- Introducing Gemma 3n: The developer guide: https://simonwillison.net/2025/Jun/26/gemma-3n/
- Mistral-Small 3.2: https://simonwillison.net/2025/Jun/20/mistral-small-32/
- Trying out the new Gemini 2.5 model family: https://simonwillison.net/2025/Jun/17/gemini-2-5/
- o3-pro: https://simonwillison.net/2025/Jun/10/o3-pro/
- Magistral — the first reasoning model by Mistral AI: https://simonwillison.net/2025/Jun/10/magistral/
- The last six months in LLMs, illustrated by pelicans on bicycles: https://simonwillison.net/2025/Jun/6/six-months-in-llms/
- Live blog: Claude 4 launch at Code with Claude: https://simonwillison.net/2025/May/22/code-with-claude-live-blog/
- Gemini 2.5: Our most intelligent models are getting even better: https://simonwillison.net/2025/May/20/gemini-25/
- Tucked into today's Google I/O keynote, a blink-and-you'll miss it moment : The pelican in the keynote was created by Al: https://simonwillison.net/2025/May/20/google-io-pelican/
- In addition to my workshop the other day I'm also participating in the poster session at PyCon US this year. This means: https://simonwillison.net/2025/May/17/pycon-poster/
- Gemini 2.5 Pro Preview: even better coding performance: https://simonwillison.net/2025/May/6/gemini-25-pro-preview/
- Qwen 3 offers a case study in how to effectively release a model: https://simonwillison.net/2025/Apr/29/qwen-3/
- Start building with Gemini 2.5 Flash: https://simonwillison.net/2025/Apr/17/start-building-with-gemini-25-flash/
- GPT-4.1: Three new million token input models from OpenAI, including their cheapest model yet: https://simonwillison.net/2025/Apr/14/gpt-4-1/
- Putting Gemini 2.5 Pro through its paces: https://simonwillison.net/2025/Mar/25/gemini/
- deepseek-ai/DeepSeek-V3-0324: https://simonwillison.net/2025/Mar/24/deepseek/
- OpenAI platform: o1-pro: https://simonwillison.net/2025/Mar/19/o1-pro/
- mlx-community/OLMo-2-0325-32B-Instruct-4bit: https://simonwillison.net/2025/Mar/16/olmo2/
- Notes on Google’s Gemma 3: https://simonwillison.net/2025/Mar/12/gemma-3/
- Initial impressions of GPT-4.5: https://simonwillison.net/2025/Feb/27/introducing-gpt-45/
- Claude 3.7 Sonnet and Claude Code: https://simonwillison.net/2025/Feb/24/claude-37-sonnet-and-claude-code/
- Andrej Karpathy’s initial impressions of Grok 3: https://simonwillison.net/2025/Feb/18/andrej-karpathy-grok-3/
- Gemini 2.0 is now available to everyone: https://simonwillison.net/2025/Feb/5/gemini-2/
- DeepSeek-R1 and exploring DeepSeek-R1-Distill-Llama-8B: https://simonwillison.net/2025/Jan/20/deepseek-r1/
- microsoft/phi-4: https://simonwillison.net/2025/Jan/8/phi-4/
- Gemini 2.0 Flash “Thinking mode”: https://simonwillison.net/2024/Dec/19/gemini-thinking-mode/
- Veo 2: https://simonwillison.net/2024/Dec/16/veo-2/
- <model-viewer> Web Component by Google: https://simonwillison.net/2024/Dec/13/model-viewer/
- Sora: https://simonwillison.net/2024/Dec/9/sora/
- I can now run a GPT-4 class model on my laptop: https://simonwillison.net/2024/Dec/9/llama-33-70b/
- New Gemini model: gemini-exp-1206: https://simonwillison.net/2024/Dec/6/gemini-exp-1206/
- QwQ: Reflect Deeply on the Boundaries of the Unknown: https://simonwillison.net/2024/Nov/27/qwq/
- Say hello to gemini-exp-1121: https://simonwillison.net/2024/Nov/22/gemini-exp-1121/
- Recraft V3: https://simonwillison.net/2024/Nov/15/recraft-v3/
- Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac: https://simonwillison.net/2024/Nov/12/qwen25-coder/
- Pelicans on a bicycle: https://simonwillison.net/2024/Oct/25/pelicans-on-a-bicycle/