Breaking

Independent testers prefer GPT-6 Astra over Claude Fable 5.1 on a Porsche 911 GT3 RS Blender job

September 7, 2026 at 14:32 EDT

Foundation Models
Media Generation
AI Agents

Independent testers put OpenAI's GPT-6 Astra and Anthropic's Claude Fable 5.1 on the same Porsche 911 GT3 RS modeling job in Blender and reported nearly identical runtimes with a modest bill gap: $132.03 and 2.58 hours for Astra versus $114.41 and 2.56 hours for Fable 5.1. They judged Astra's mesh more detailed and closer to the production car, describing Fable 5.1's result as blocky.

Sep 7, 2026 · OpenAI vs Anthropic

Same clock. Different car.

On a Porsche 911 GT3 RS built in Blender, GPT-6 Astra and Claude Fable 5.1 finished in about 2.6 hours. Testers called Astra closer to the real car for about $18 more.

$132.03

Astra cost

$114.41

Fable 5.1 cost

2.58 h

Astra runtime

2.56 h

Fable runtime

Cost on the same Blender job

Astra spent about 15% more

$132.03

GPT-6 Astra

$114.41

Claude Fable 5.1

Each block is about $10. Runtime was effectively tied: 2.58 hours vs 2.56 hours.

Where Astra led

Closer Porsche mesh in the Blender test, plus OpenAI's BenchCAD score of 95.9% vs 84.3% and a 19-of-20 robot-arm bowl completion vs 8 of 20.

Where Fable held

Lower bill on this run ($114.41). Some developers still prefer Fable 5.1 for code quality, and it leads the Intelligence Index at 65.7 vs 61.2.

Shared sticker, different bill

Both flagships list at $10 input / $50 output per 1M tokens. Cache and long-context rules still split the real cost.

Cache reads · Astra $1.00 · Fable $0.25 This job · Astra $17.62 more Access · Fable GA · Astra still rolling out

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.