Unverified50% confidenceBenchmarkExact time
Qwen 3.8 has a base speed of 22 tokens/s on the M5 Max, which can be boosted to 34 or even 56 tokens/s with MTP enabled.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/10/2026
First Seen
Valid until: 12/9/2026
Sources
Related Entities
Related Claims
UnverifiedQwen3.7 Max already supports context lengths at the million-token level.70% similarUnverifiedIn token generation tests for the Gemma 4 12B model, Spark reached 26.4 tokens/s, Halo 24.6 tokens/s, and M4 Pro was the fastest at 33.8 tokens/s.70% similarUnverifiedRunning quantized models from the Qwen3 series on a Mac M3 Ultra, inference speed reaches approximately 34 tokens/s under Q8 quantization.69% similarUnverifiedQwen2.5-Max-0902 has a maximum output of 130,000 tokens, with the limit rising to 260,000 tokens in reasoning mode.68% similarUnverifiedThe Qwen3.6 34B 80-layer model achieves an MTP speculative decoding acceptance rate of approximately 78% on LM Studio, while the original 27B achieves only approximately 57–58%67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/887804API
curl https://kongchang.com/api/v1/knowledge/claims/887804MCP
get_claim(id=887804)