Expired50% confidenceBenchmarkExact time
In prefill speed tests, Halo achieved approximately 650 tokens/s while Spark achieved approximately 2000 tokens/s, with Spark being nearly three times faster.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/12/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/10/2026(expired)
Sources
AMD Ryzen AI Halo Hands-On: Can x86 Architecture Challenge NVIDIA's DGX Spark?
bilibili量子菠萝_7/7/2026
Related Claims
UnverifiedUnder a 4×RTX 3090 SXM4 configuration, the quantized version of GLM-5.2 achieves a token generation speed of approximately 6.3–6.8 tokens/second and a prompt evaluation speed of approximately 10.6 tokens/second.75% similarUnverifiedGrok 4.5 outputs at 80 tokens per second and features a context window of up to 500,000 tokens (approximately 400,000 English words).71% similarUnverified使用SGLang同时开启Prefix Cache和MTP后，token生成速度提升到约40-50 tokens/s70% similarUnverifiedHalo's prefill result of 650 tokens/s is nearly double the speed measured for the same chip last year, with the improvement stemming from AMD contributing code to upstream open-source projects such as llama.cpp to advance the maturity of the ROCm ecosystem.69% similarUnverifiedIn a controlled test, the 35B model ran at approximately 100 tokens/second (Mac Studio), roughly five times the 9B model's 16 tokens/second68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/488914API
curl https://kongchang.com/api/v1/knowledge/claims/488914MCP
get_claim(id=488914)