Unverified50% confidenceFactExact time
Strata is a local inference engine that can run 180B-class MoE models on consumer-grade GPUs with 12 GB of VRAM.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/5/2026
First Seen
Valid until: 1/3/2027
Sources
Related Entities
Related Claims
UnverifiedStrata is an open-source inference runtime capable of running the 125-billion-parameter Qwen3.8 Flash Next model on an RTX 5070 (12 GB VRAM) at an output speed of approximately 90 tokens per second.68% similarUnverifiedStrata offers multiple quantization tiers: 66GB Q20 (fastest), 68GB IQ2XS (balanced), 76GB IQ3XS (best quality), with the test using IQ3S at 83.6GB.64% similarVerified一个原本需要60GB显存的30B参数模型，经过量化后能在16GB甚至8GB显存的消费级显卡上运行62% similarUnverifiedOllama支持GGUF格式的量化模型，可以在消费级硬件（如16GB内存的笔记本电脑）上流畅运行7B-13B参数规模的模型61% similarVerifiedTaking a 70B parameter model with FP16 precision and 32 attention head layers as an example, the KV Cache for a single 200K-length request requires approximately 80–120 GB of VRAM, far exceeding the 80 GB VRAM limit of a single A100 GPU.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/978623API
curl https://kongchang.com/api/v1/knowledge/claims/978623MCP
get_claim(id=978623)