Unverified50% confidenceFactExact time
Gemma 3 offers multiple sizes covering different hardware, including a minimum version of approximately 1GB, medium-sized versions, and large versions such as 12B and 27B.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedThe FP16 version of Gemma 3 12B requires approximately 24 GB of VRAM, while the Q4_K_M version reduces this to approximately 7–8 GB, with inference speed improving by approximately 2–3x and structured output accuracy losing only approximately 2–5%.75% similarVerifiedGemma 4的1B/2B模型量化后仅需4GB显存即可运行71% similarUnverifiedThe Gemma 4 26B MOE version can run on a graphics card with 24GB of VRAM.69% similarUnverifiedThe officially supported context length of Gemma 3 12B is 128K tokens, but the default num_ctx in Ollama is only 2048.68% similarUnverifiedThe Gemma 4 26B MOE version supports an ultra-long context window of 256K.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/905775API
curl https://kongchang.com/api/v1/knowledge/claims/905775MCP
get_claim(id=905775)