Breaking

GPT-6 Astra and Claude Fable 5.1 Run the Same Prompts After Back-to-Back Launches

September 4, 2026 at 15:46 EDT

Foundation Models
Software Dev & Coding
AI Agents

OpenAI’s GPT-6 Astra and Anthropic’s Claude Fable 5.1, released two days apart at the same headline API price, are being run on identical prompts — and the visual results are already colliding with independent scoreboards.

September 4, 2026 · OpenAI · Anthropic

Same prompt. Two flagships.

GPT-6 Astra and Claude Fable 5.1 launched two days apart at the same API price. Identical-prompt tests and the scoreboards do not agree on a winner.

API price (both)

$10 / $50

per 1M in / out tokens

Intelligence Index

61 vs 66

Astra vs Fable 5.1

FrontierMath T4

97.6%

Astra vs 87.8% Fable

AutomationBench

41.4%

Astra vs 31.4% Fable

AutomationBench, drawn to scale

Business-workflow completion on the same unit. Column height is proportional (8 blocks vs 6).

41.4%

GPT-6 Astra

31.4%

Claude Fable 5.1

Cache reads: 4× price gap

One block per $0.25 per million tokens. Fable 5.1 is one-quarter the Astra cache-read rate.

$0.25

Fable 5.1 cache

$1.00

Astra cache

Takeaway: On cache-heavy agent loops, Fable 5.1 is the cheaper of two $10 / $50 models.

Where Astra lands

On matching Blender and game prompts, testers often preferred Astra’s finished lighting, furniture, and geometry. It also leads OpenAI’s table on FrontierMath, AutomationBench, and BenchCAD.

Where Fable holds

Fable 5.1 still leads the Artificial Analysis Intelligence Index (66 vs 61) and Coding Agent Index (70 vs 67). Some testers still prefer its visual taste and game feel.

The split

No single winner on the same prompt

Astra is often stronger at finished 3D, math, and desktop automation. Fable 5.1 still leads some aggregate intelligence and coding-agent scores and remains cheaper on cached agent work. Access is phased; ARC-AGI-3’s 99.9% score uses a provider adapter, versus 62.7% on the standard harness.

Continue reading

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