Kimi K3 Might Not Actually Save Money
theo · x · 2026-07-18
Theo pointed out that the key takeaway about Kimi K3 isn't that it's "cheap," but that its total real-world usage cost is roughly similar to GPT-5.6 Sol.
- K3's per-token price is half that of GPT-5.6 Sol, but GPT-5.6 typically uses only about half the tokens.
- Furthermore, since GPT-5.6's TPS is about 2x faster, it completes tasks roughly 4x faster overall, making them comparable in terms of "cost per completed task."
- He still acknowledges K3's capabilities and will continue to use it extensively, but objects to it being simply marketed as noticeably cheaper. The attached chart illustrates this using a stacked cost graph.
More from Models
- Week 3 of "Fable 5.5 launches this week" hype circles: no new frontier models this October — daniel_mac8 · 2026-10-27
- 125 poisoned samples backdoored Qwen2.5-7B — and open-weight supply chain risks debated — lxrjl · 2026-10-27
- Speculative leaderboard shows "Gemini 4 Argon" vs "GPT-6.1 Sol"; observer calls MiniMax parity suspicious — teortaxesTex · 2026-10-27
- Cohere sandbox: Qwen model tried to upload probe file to Beijing cloud server unprompted — simonguozirui · 2026-10-27
- Kimi K3.1 Pro tops open-weight models on FrontierML benchmark, still far behind Opus 5.5 — teortaxesTex · 2026-10-27
- Across 7 tasks, highest reasoning setting isn't the best for most open models — keunwoochoi · 2026-10-27