AI/TLDR

Fireship · 2026-09-09 · notable

Fireship — 'I built the same game with Astra and Fable 5.1'

Fireship gives GPT-6 Astra and Claude Fable 5.1 the same brief — build a game — and compares what each one produced. The video went up on 9 September 2026, now that Astra is generally available.

Fireship video thumbnail comparing GPT-6 Astra and Claude Fable 5.1 on the same game

Fireship builds the same game twice — once with GPT-6 Astra, once with Claude Fable 5.1 — and compares the results.

What is it?

'I built the same game with Astra and Fable 5.1... only one was fun' went up on the Fireship channel on 9 September 2026. The episode takes GPT-6 Astra, which has only just become generally available, alongside Claude Fable 5.1, and gives each the same job. Fireship sets up the test in the description as 'now that Astra is actually available to use, let's see if it lives up to the hype.'

How does it work?

The format is a side-by-side build rather than a benchmark run. Each model gets the same brief and the same target, and Fireship judges the output the way a player would — whether the finished game is actually fun — instead of scoring it on a coding eval. That puts two flagship coding models against a task where taste, not test pass rate, settles it.

Why does it matter?

GPT-6 Astra and Claude Fable 5.1 both shipped in the first days of September 2026, and developers are choosing between them right now on benchmark tables and word of mouth. A head-to-head on one concrete build is the comparison that travels fastest: the video passed 117,000 views in its first three hours on the channel.

Who is it for?

developers choosing between Astra and Fable 5.1

Sources · 3 outlets

Tags

  • video
  • fireship
  • gpt-6-astra
  • claude-fable-5-1
  • coding-agents
  • model-comparison
  • game-development
  • explainer

← All releases · Learn AI