Local AI Advisor

Can Surface Laptop Ultra · RTX Spark N1X (20-core) · 128 GB run Muse Glimmer (30B)?

128 GB unified · 300 GB/s memory bandwidth · Muse Glimmer (30B) has 29.6B parameters.

Short answer

Yes, acceptable for chat; borderline for agentic coding.

Best quantization for chat: Q4_K_M (16 GB of weights) on Ollama. Generation: 9–15 tok/s, leaving about 99 GB free for your other apps.

How it rates for each workload

Same model and machine, different jobs. Agents and long documents send far bigger prompts than chat, so they need much more speed and memory.

WorkloadVerdictQuantGenerationFirst replyContext
Chatwith Open WebUIAcceptableQ4_K_M9–15 tok/sunder 1 s8KDetails
Coding questionswith ContinueAcceptableQ4_K_M9–15 tok/s≈3 s16KDetails
Coding in a repositorywith AiderAcceptableQ4_K_M9–15 tok/s≈9 s32KDetails
Agentic codingwith OpenCodeBorderlineQ4_K_M9–14 tok/s≈10 s32KDetails
Long documents (64K)with Open WebUIAcceptableQ4_K_M8–14 tok/s≈35 s64KDetails

Speeds are bandwidth-based estimates; no direct benchmark exists for this pair. “First reply” is the time to read the workload’s opening prompt. How we calculate

Different tool, context size or workload?

These pages use typical settings. The full check models your exact setup.