Expired50% confidenceFactExact time
WisprGemma requires downloading approximately 3.5GB of model weights on first load.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/11/2026
First Seen
Valid until: 10/9/2026(expired)
Sources
WisprGemma：纯浏览器本地语音输入，数据不上云
redditr/GoogleGeminiAI7/10/2026
Related Claims
UnverifiedLlama 3.1 7B量化后仅需约4-5GB显存即可运行67% similarUnverifiedOn a 16 GB M4 Mac Mini, after loading Ornith 9B GGUF, 6.2 GB is used for model weights, and the maximum context length can reach 260,000 tokens.64% similarUnverifiedThe Qwen 3.5 9B model is approximately 6.6 GB and is suited for GPUs with 8 GB of VRAM64% similarUnverifiedFull fine-tuning of Llama-3 (8B) requires approximately 80 GB or more of VRAM, while Llama-3 (70B) requires more than 400 GB.64% similarUnverifiedAfter 4-bit quantization, Llama 3.3 70B has a model size of approximately 40 GB and can run on a Mac Studio with 64 GB of memory or a single professional GPU with 48 GB of VRAM.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/488226API
curl https://kongchang.com/api/v1/knowledge/claims/488226MCP
get_claim(id=488226)