Expired50% confidenceBenchmarkExact time
Strata's chunked prefill can reach prefill speeds of 1000 to 2600 Token/s on the graphics card.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
11/4/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 11/18/2026(expired)
Sources
Running 125B MoE on a 12GB GPU? Inside Strata's Three-Tier Cooperative Engine
bilibiliNicksxs11/3/2026
Related Entities
Related Claims
UnverifiedWith a configuration of 16384 Tokens per card, 7168 values per Token (2 bytes each), and 6 copies, a single Dispatch requires approximately 1.4 GB of data movement per card, and FP8 can halve this amount.68% similarUnverifiedOn consumer GPUs with 12GB to 16GB of VRAM, Strata can achieve generation speeds of 40 to 95 Token/s when running a 125B MoE model.65% similarUnverifiedOfficial estimates suggest a 24GB GPU can run approximately 100–140 tokens/second.64% similarUnverifiedWith an approximately 1.6M ultra-long context and 128 concurrent requests, throughput can reach approximately 3,700 tokens/s (aggregated).63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1134852API
curl https://kongchang.com/api/v1/knowledge/claims/1134852MCP
get_claim(id=1134852)