Model comparison

Claude Fable 5.1 vs Muse Spark 1.3

Compare pricing, context windows and capabilities from OpenRouter. Open either model’s page to check availability and try it online.

anthropicClaude Fable 5.1newanthropic/claude-fable-5.1metaMuse Spark 1.3newmeta/muse-spark-1.3

Measured intelligence and task cost

Two current intelligence leaders from different providers. Artificial Analysis index v4.3, updated 2026-09-21.

Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback): 53.4 intelligence · $7.63 / AA task

Muse Spark 1.3 (max): 48.1 intelligence · $1.60 / AA task

Highest reasoning effort with both measurements available. Benchmark task costs are not OpenCraft task quotes; your settings and workload can differ.

Input price /M$10.00$1.258× cheaper

Output price /M$50.00$4.2512× cheaper

Cache read /M$0.25$0.151.7× cheaper

Context window1M1M

Catalog date1 Sept 20262 Sept 2026

What the same turn costs

Estimates from published token rates — stated token counts, not measurements.

Quick answer2K tokens in, 600 outClaude Fable 5.1$0.05Muse Spark 1.3$0.00509.9× cheaper

10-slide deck build60K tokens in, 9K outClaude Fable 5.1$1.05Muse Spark 1.3$0.1139.3× cheaper

150-page PDF read200K tokens in, 4K outClaude Fable 5.1$2.20Muse Spark 1.3$0.2678.2× cheaper

Capabilities

Bothtool-usevisionreasoning

Claude Fable 5.1: text, image, file → text · Muse Spark 1.3: text, image, video, file, audio → text

Which model should you use?

Start with the token cost for your workload, the context you need, and supported input types. A lower price or larger context window does not establish better answer quality.

To compare answers, open each model in a separate chat and use the same prompt. Token tables use published specs and estimates; any measured AA results above are a separate benchmark.

Try Claude Fable 5.1 — pricing and availability Try Muse Spark 1.3 — pricing and availability