<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Coding-Assistant on Best of AI</title><link>https://bestofai.io/tags/coding-assistant/</link><description>Recent content in Coding-Assistant on Best of AI</description><generator>Hugo</generator><language>en-US</language><lastBuildDate>Sun, 09 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://bestofai.io/tags/coding-assistant/index.xml" rel="self" type="application/rss+xml"/><item><title>Muse Spark 1.2</title><link>https://bestofai.io/models/muse-spark-1-2/</link><pubDate>Sun, 09 Aug 2026 00:00:00 +0000</pubDate><guid>https://bestofai.io/models/muse-spark-1-2/</guid><description>&lt;p&gt;Muse Spark 1.2 is a coding-focused point release of Muse Spark 1.1 rather than a new base model, shipped August 5, 2026 alongside Muse Code, Meta&amp;rsquo;s first terminal coding agent. It keeps the 1,048,576 token context window and roughly 131,072 max output tokens from 1.1, and was co-trained with the new agent using rejection-sampled harness trajectories plus a self-improvement loop. Meta reports a jump from 76.2 to 82.9 percent on Terminal-Bench 2.1 and from 53.0 to 59.3 percent on DeepSWE v1.1 over the previous version. Standard API pricing holds at $1.25 per million input tokens and $4.25 per million output tokens, and Meta added a separate contributor tier at $0.10 input and $0.20 output in exchange for permission to train future models on submitted prompts and completions.&lt;/p&gt;</description></item></channel></rss>