$ diff hy4-preview kimi-k3
Hy4 preview vs Kimi K3
Values come from Hy4 preview and Kimi K3, where each is cited to its source. This page states no benchmark result and ranks neither model — it puts two published specifications next to each other. Where a lab has not published a figure, the row says so rather than guessing.
TencentHy4 preview
- context
- 1M
- weights
- open
- $/M in
- $0.834
- $/M out
- $2.501
- context
- 1M
- weights
- open
- $/M in
- $3
- $/M out
- $15
What actually differs
- Context window
- Identical — both accept 1M tokens, so context length is not a reason to pick either.
- Input price
- Hy4 preview at $0.834/M against $3/M — 3.6× cheaper to feed. Standard rates: one of these labs also quotes a lower cached-input tier, which applies only when a prefix is reused — see the full spec below.
- Output price
- Hy4 preview at $2.501/M against $15/M — 6.0× cheaper to generate. Output dominates the bill on most agentic workloads, where the model writes far more than it reads.
- Weights
- Both publish weights — Hy4 preview under Apache-2.0, Kimi K3 under Modified MIT (commercial use permitted; weights publicly downloadable). Check the licences rather than assuming they permit the same commercial use.
- Recency
- Hy4 preview shipped 43 days after Kimi K3 (2026-08-28 vs 2026-07-16).
Full spec
| Attribute | Hy4 preview | Kimi K3 |
|---|---|---|
| Developer | Tencent | Moonshot AI |
| Released | 2026-08-28 | 2026-07-16 |
| Context window | over 1,000,000 | 1M tokens |
| Pricing | $0.834/M input · $2.501/M output | $0.30/M cache-hit input · $3.00/M cache-miss input · $15.00/M output |
| License | Apache-2.0 | Modified MIT (commercial use permitted; weights publicly downloadable) |
| Availability | Hugging Face (BF16 + FP8), ModelScope, GitCode, CNB, Tencent Cloud TokenHub, OpenRouter, WorkBuddy, CodeBuddy, Yuanbao, ima | Kimi Code, Kimi app, Kimi API; weights on Hugging Face (`moonshot-ai/kimi-k3`, released 2026-07-27) |