What was the original game concept, and how did the two AI versions differ?

The original 2022 GPT-3 premise described a team heist game with thieving raccoons robbing banks and stealing art. Claude Fable 5 simplified it to a single raccoon collecting coins and fish in a backyard; GPT-5.6 Sol stayed closer to the heist theme, placing three raccoons in a museum where they must stack on top of each other to steal a golden sardine from its case.

What bug did the GPT-5.6 Sol version have, and how was it fixed?

Each raccoon character had a giant black sphere floating over its head where an eyeball should have been. The developer fixed it by asking Codex directly why the raccoons had huge spheres, then requesting a fix, which the AI applied successfully.

How long did the project take, and where can I see the code?

Codex spent 52 minutes on the project. The GitHub repository includes the game code, textures, image-generation prompts, and the full Codex transcript.

Large Language ModelsAI Coding AssistantsSimon Willison's WeblogAug 7, 2026Aug 8, 2026, 06:00 JST

GPT-5.6 Sol Ultra builds better game than Claude in test

Send on LINE

3 Key Points

What happened
On August 7, 2026, a developer tested GPT-5.6 Sol Ultra (running in Codex Desktop with aggressive sub-agent use) on a raccoon heist game premise originally generated with GPT-3 four years ago. The resulting game significantly improved on Claude Fable 5's version—moving from a single raccoon collecting coins in a backyard to a museum heist where three raccoons stack to steal a golden sardine.
Why it matters
The test shows GPT-5.6 Sol Ultra's ability to produce more complex, feature-rich game output from the same prompt. Though the first version had a bug (oversized eyeballs on each raccoon), the AI failed to spot it during development—requiring manual fixes via follow-up prompts. This suggests that while reasoning power improved, systematic review and independent problem detection remain gaps.
What to watch
Codex spent 52 minutes on the full project; the estimated cost at full API pricing (versus a monthly subscription used here) is available in the AgentsView cost estimate shared in the GitHub repository. The complete transcript and game code, including textures and image-generation prompts, are available on GitHub.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

The test represents a direct head-to-head comparison of two AI systems on an identical creative task: building a playable game from a single natural-language prompt. The original 2022 premise—generated by GPT-3—was deliberately reused four years later to measure progress. GPT-5.6 Sol's output demonstrated ambition and structural improvement over Claude Fable 5, preserving the heist concept and adding the emergent mechanic of stacking characters to achieve a goal, whereas Claude's version simplified the premise into a single-character coin-collection game. However, the raccoon eyeball bug reveals a practical limitation: despite the AI reviewing screenshots during development, it failed to detect and self-correct an obvious visual glitch. The developer's manual fix—which required explicit prompts naming the problem—suggests that even aggressive sub-agent use in Codex does not automatically confer visual verification or self-critique capabilities. The cost estimate and 52-minute session length provide benchmarks for computational load, though the true API expense is masked by the developer's monthly subscription model.

FAQ

What was the original game concept, and how did the two AI versions differ?: The original 2022 GPT-3 premise described a team heist game with thieving raccoons robbing banks and stealing art. Claude Fable 5 simplified it to a single raccoon collecting coins and fish in a backyard; GPT-5.6 Sol stayed closer to the heist theme, placing three raccoons in a museum where they must stack on top of each other to steal a golden sardine from its case.
What bug did the GPT-5.6 Sol version have, and how was it fixed?: Each raccoon character had a giant black sphere floating over its head where an eyeball should have been. The developer fixed it by asking Codex directly why the raccoons had huge spheres, then requesting a fix, which the AI applied successfully.
How long did the project take, and where can I see the code?: Codex spent 52 minutes on the project. The GitHub repository includes the game code, textures, image-generation prompts, and the full Codex transcript.

Simon Willison's WeblogRead Original Article

Send on LINE

Get the latest Large Language Models news every morning

For example, today's edition would include:

Instinct raises $1B at $10B valuation for personal AI agentSiliconANGLE AI · 30m ago
Okta's Wylie: agent security needs shared safeguardsSiliconANGLE AI · 30m ago
CoreWeave's top three customers drive 70% of revenue, Vellante saysSiliconANGLE AI · 30m ago

AI-summarized, only the topics you pick: one digest a day via Email, LINE, or Slack.

Subscribe free →

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.