Unverified50% confidenceBenchmarkExact time
In testing, an RTX 5070 running the 2-bit quantized version outputs 94 tokens per second.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/16/2026
First Seen
Valid until: 10/30/2026
Sources
12GB GPU Runs a 125-Billion-Parameter LLM Locally: Strata Hands-On Test
bilibili宇智波刘能king10/15/2026
Related Entities
Related Claims
UnverifiedAccording to the README, the 12 GB RTX 5070 achieves approximately 94 tokens/second, while the 32 GB RTX 5090 can exceed 250 tokens/second78% similarVerifiedOn the RTX 5090, Muse Glimmer's generation speed increases from 75 tokens per second to 233 tokens per second, a speedup of approximately 3.1×.77% similarVerifiedIn an RTX 5070 (12GB VRAM) configuration, the local Qwen3 9B model's output speed reached approximately 70–76 tokens/second.76% similarUnverifiedOn an RTX 5090, Muse Glimmer runs at 75 tokens/second without speculative decoding, and reaches 233 tokens/second when paired with a speculative decoding model72% similarUnverifiedWhen batching 8 requests together, a single RTX PRO 6000 can achieve an aggregate throughput of approximately 220 output tok/s, whereas the old configuration of two RTX 3090s achieved only approximately 19 tok/s.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1044921API
curl https://kongchang.com/api/v1/knowledge/claims/1044921MCP
get_claim(id=1044921)