Trending:
OpenAI admin controls GitHub Issues agent approvals Gemini in Chrome UK expansion Developer tools coverage
TECHSTAGED Better software. Smarter business.

Junie Local arrives on NVIDIA RTX Spark: what the verified evidence shows beyond the announcement

By TechStaged

· 2 min read

Prepared with AI assistance from source material. Automated editorial review recorded . Automated checks can miss errors and do not represent human approval. Editorial policy and corrections.

Junie Local running on NVIDIA RTX Spark hardware
TechStaged-owned

Summary

  • Junie Local is supported on the NVIDIA RTX Spark from day one; Windows support moved from experimental to stable.
  • RTX Spark pairs a Blackwell RTX GPU with a Grace CPU to run larger local AI models in slim laptops and compact desktops.
  • Junie Local can be installed via /local and runs as a full coding agent that reads code, edits files, and runs tests; no tokens, no quotas, code stays on the machine.
  • In testing on the RTX Spark, Junie Local reached generation speeds of up to 55 tokens per second.

The JetBrains blog confirms Junie Local is supported on the NVIDIA RTX Spark from day one, with Windows support moving from experimental to stable.

RTX Spark combines a Blackwell RTX GPU with a Grace CPU to run larger local AI models in compact systems, and Junie Local leverages this hardware to enable local AI coding in everyday workflows.

WHAT JUNIE LOCAL DOES ON RTX SPARK

Installing Junie and running the /local command provides a full coding agent that can read code, edit files, and run tests, all on the local machine with no token charges and no external data leaving the device.

PERFORMANCE CHARACTERISTICS AND INFERENCE APPROACH

In testing on the RTX Spark, Junie Local reached generation speeds of up to 55 tokens per second.

Junie Local sizes its context window to fit the memory available to the GPU in your configuration, and uses Blackwell Tensor Cores for FP8 computation along with speculative decoding to accelerate generation.

WINDOWS SUPPORT, HARDWARE REQUIREMENTS, AND INSTALLER FLOW

Junie Local is now stable on Windows and supports NVIDIA RTX GPUs with 24 GB of VRAM, including cards like the RTX 3090, 4090, or 32 GB 5090, on 64-bit Windows.

The installer checks your PC, verifies the GPU and driver, then installs an optimized inference engine, the model, and the Junie integration.

MODELS AND SETUP OPTIONS

Users can choose between Qwen3.6-27B and Qwen3.8-3.6-27B-blend, a merged model that blends coding abilities. More models are expected in the future.

WHAT THIS MEANS FOR DEVELOPERS AND WHAT’S NEXT

The evidence suggests Junie Local on RTX Spark could broaden access to local AI coding on capable Windows machines with 24 GB VRAM, while ongoing work aims to improve model formats and inference speed.

NVIDIA and Junie researchers indicate that additional optimizations and benchmarks are forthcoming.

SOURCES

Developer Tools Open Source License Compliance GA for GitHub Enterprise Cloud signals enterprise-wide governance • 2 min read Developer Tools Governed, agentic code review on GitLab Duo Agent Platform: what the verified evidence says about feasibility, cost, and governance • 4 min read Developer Tools Void’s Open-Source Stack: What Open Tools Platform Could Mean for Full-Stack Devs • 2 min read Developer Tools Rust 1.99.1 arrives with targeted fixes for wcslen, FFI, and macOS rust-lld quirks: an evidence-based read • 3 min read Developer Tools GitLab's AI Governance Dashboard beta: a single view of agent activity • 2 min read Developer Tools GitLab’s Business Logic Security Analyzer: what verified evidence reveals beyond the announcement • 2 min read Security Agents are your newest insiders: Security must move at their speed • 2 min read Developer Tools Thin Orchestrator, Deep Specialists: How Meta’s 1,300-Line Router Strategy Aims to Make AI Agents More Reliable • 3 min read