Unverified50% confidenceFactExact time

The experiment uses an n1-standard-2 VM, a single passively cooled Tesla T4, driver 615.71.09, vLLM 0.29.0, with the model being Gemma 4 E2B at 4-bit weights, single request, greedy decoding, 256 tokens per output.

1

Sources

50%

Confidence

Short-term (~14 days)

Relevance

10/16/2026

First Seen

Valid until: 10/30/2026

Sources

Cloud T4 Falls Short of Spec: Gemma 4 Decoding Slows 7% Due to Heat

Related Entities

n1-standard-2product NVIDIA T4product vLLM 0.29.0product Gemma 4 E2Bmodel

Related Claims

UnverifiedThe Tesla T4 is an NVIDIA data center inference card based on the Turing architecture, with a single-card TDP of 70W and a passive cooling design69% similar UnverifiedThe Tesla V100 has a compute capability of 7.0.68% similar UnverifiedSiliconFlow has launched the SIM 4.0.29 BA4B model with free API access; the model natively supports a 256K context window and specializes in complex engineering and long-horizon agent tasks.65% similar UnverifiedVKORE's benchmark test uses Google's open-source Gemma 4 26B model, running on an eight-card, 32-chip Korean NPU server.65% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/1047194

API

curl https://kongchang.com/api/v1/knowledge/claims/1047194

MCP

get_claim(id=1047194)