Unverified50% confidenceBenchmarkExact time
In Ollama's library, qwen3.5:0.8b using the MLX engine's mxfp8 quantization achieved 2,264.6 tok/s prompt processing and 95.3 tok/s generation.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/15/2026
First Seen
Valid until: 1/13/2027
Sources
Related Entities
Related Claims
UnverifiedGFX1151 engine has been deeply optimized for running the Qwen3 series of large models on the AMD AI MAX+ 395 platform67% similarUnverifiedOn 8 AMD Instinct MI355X accelerator cards, using the vLLM inference framework, the GLM-5.3 model achieved a single-user decoding speed of 469 tok/s.66% similarUnverifiedFA4 MX8 achieves performance of 2.85 PF/s for the forward pass and 2 PF/s for the backward pass on typical LLM shapes.64% similarUnverifiedLFM2-VL achieves a 1.66x to 2.56x speedup at batch 1 via DSpark speculative decoding.63% similarUnverifiedQwen3 Coder Next is an 80B-parameter mixture-of-experts (MoE) coding model, deployed by vLLM on a single A100 80GB, using the 4-bit AWQ quantized version published by cyankiwi.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1039234API
curl https://kongchang.com/api/v1/knowledge/claims/1039234MCP
get_claim(id=1039234)