Muse Spark 1.2 vs 1.3: four index points, same token rates

Muse Spark 1.3 xhigh scores 61 versus 57 for 1.2 xhigh, with unchanged standard input and output token prices.

Create account & add credits Run this with Muse Spark 1.3

Last verified 3 September 2026.

Use Muse Spark 1.3 in 4 steps

Step 1 of 4

Try Muse Spark 1.3 in the OneInfer playground without writing code, or create an API key in the Console to call it from your app.

export ONEINFER_API_KEY="your-api-key"

Version comparison

Metric	1.2	1.3
Release	5 August 2026	2 September 2026
AA index (xhigh)	57	61
Input / 1M tokens	$1.25	$1.25
Output / 1M tokens	$4.25	$4.25
Weights (3 September 2026)	Closed	Closed

Upgrade test

Replay representative prompts with the same system instructions and tool permissions. Compare correctness, tokens consumed, retries and user-visible latency before switching traffic.

Ready to test the workflow?

Create account & add credits

Efficiency is workload-dependent

Meta's fewer-token coding claim and Artificial Analysis's higher-token index result come from different workloads. Neither guarantees lower production cost.

Frequently asked questions

How current is this information?

This page reflects sources checked on 3 September 2026. Verify current access, pricing and provider terms before an evaluation.

Put Muse Spark 1.3 to work

Fund a controlled evaluation, test it on your own inputs, and measure quality and cost on your workload.

Create account & add credits Run this with Muse Spark 1.3