Unverified50% confidenceBenchmarkExact time
TerraGemma primarily uses the gemma3:1b model, which is approximately 815 MB and completes inference in approximately 500 ms on an ordinary laptop CPU.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
UnverifiedThe model gemma4:e2b-it-qat runs locally via Ollama on a CPU-only i5-1135G7 laptop, with a download size of 4.3 GB, a cold-start time of approximately 14 seconds, a warm-start time of 3.6 to 4.0 seconds, and a speed of approximately 15 to 20 tokens/second.71% similarUnverifiedA sandbox engine with 8 million parameters (8.3M) achieves a decoding speed of approximately 2,000 tokens/second on a 6-core desktop CPU released in 2019.66% similarUnverifiedApple M2/M3 chips adopt a Unified Memory Architecture (UMA), and llama.cpp can achieve approximately 30–50 tokens/second inference speed for 7B models via the Metal API.64% similarUnverifiedThe core device of this local AI solution is equipped with an NVIDIA GB-series (Grace Blackwell architecture) superchip, featuring 128GB unified memory, a 20-core ARM CPU, and 4TB of storage, priced at approximately $7,000.63% similarUnverifiedThe 4-bit quantized GGUF format version of llama.cpp enables the Llama 3.1 70B model to run on an Apple M2 Ultra with 64GB unified memory at an inference speed of approximately 20–30 tokens/second.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1003604API
curl https://kongchang.com/api/v1/knowledge/claims/1003604MCP
get_claim(id=1003604)