Hermes time-to-first-token test: V4.1 Flash second at 0.91s median

On short prompts, MiniMax-M3.1-Flash led at 0.89s with DeepSeek V4.1 Flash at 0.91s, while GLM, Kimi and GPT-6.1-sol all exceeded 4s.

Hobson

@yafeishiArchived 2 days agoOriginally posted 2026.10.27

Hermes time-to-first-token test: V4.1 Flash second at 0.91s median (image 1 of 2)

Hermes time-to-first-token test: V4.1 Flash second at 0.91s median (image 2 of 2)

Hobson @yafeishiOriginally posted 2026.10.27

Open original on X

Hermes 首字延迟实测（原生 gateway / Magpie / 低推理 / 短提示×3）中位： MiniMax-M3.1-Flash 0.89s DeepSeek V4.1 Flash 0.91s Qwen3.8 Flash 1.32s GLM-5.3 Flash 4.12s Kimi k3.1-pro 4.15s GPT-6.1-sol 4.79s 小样本短提示，不是质量榜。

x.com/yafeishi/status/2114760045342675272Archived 2026.10.27 12:10

Open original on X