Kimi K3.1 Pro Highspeed vs Pro: we measured the real speedup

Описание и применение

Highspeed ускоряет генерацию токенов в 2–3 раза, но не ускоряет первый токен. Общая скорость выполнения зависит от формы запроса: 1,7× для коротких ответов, 2× для длинных кодовых ответов и 1,25× для больших входов. Стоимость за запрос почти вдвое выше, но при работе с агентами экономия времени стоит 11 центов за 4 минуты.

Задача

Determine whether the higher‑priced Highspeed endpoint delivers a worthwhile speed‑to‑cost advantage over the standard Kimi K3.1 Pro model.

Результат

Highspeed achieves 2–3× faster token generation but does not reduce first‑token latency. End‑to‑end speedup is 1.7× for short answers, 2× for long code answers, and 1.25× for large inputs. Cost per request is roughly double, though agent runs incur only 1.5× higher cost due to prompt caching.

Рабочий процесс

• 1. Send 48 streamed /v1/chat/completions requests (short answer, long code answer, 52K‑token log) to each model, 8 runs per type.
• 2. Measure first‑token time, total time, output tokens/s, and cost per request.
• 3. Run OpenCode 1.19.4 coding agent on six Python tasks twice per model, recording time and cost per run.
• 4. Aggregate results and compute speedup and cost ratios.
• 5. Analyze trade‑offs for interactive vs batch workloads.

Доказательства

Highspeed generated text 2 to 3 times faster.

Источник

Both models found the root‑cause line in all 16 long‑input runs and quoted it exactly.

Источник

Highspeed cost 1.5 times as much, not 2.

Источник