Claude Sonnet 5 vs Kimi K2.7 Code
Both are Mid-tier models, from Anthropic and Moonshot AI respectively. Kimi K2.7 Code is the cheaper choice overall at a 4:1 input:output ratio.
Which is cheaper
- Input price: Kimi K2.7 Code is cheaper ($0.95 vs $2.00).
- Output price: Kimi K2.7 Code is cheaper ($4.00 vs $10.00).
- Blended, 4:1 input:output ratio: Kimi K2.7 Code is 56.7% cheaper ($1.56 vs $3.60 per weighted 1M tokens).
Side by side
| Spec | Claude Sonnet 5 | Kimi K2.7 Code |
|---|---|---|
| Provider | Anthropic | Moonshot AI |
| Tier | Mid | Mid |
| Context window | 1M | 262K |
| Max output | 128K | — |
| Input $/1M | $2.00 | $0.95 |
| Output $/1M | $10.00 | $4.00 |
| Cache read $/1M | $0.20 | $0.19 |
| Cache write $/1M | $2.50 | — |
| Modalities | text, image | text |
Claude Sonnet 5. 5-minute cache write shown; 1-hour cache write is $4/MTok. The page now states explicitly the previously-scheduled Sep 1 2026 increase to $3/$15 will NOT occur - $2/$10 is permanent.
Kimi K2.7 Code. Coding-focused model; input is cache-miss price, cache_read is cache-hit price ($0.19). A HighSpeed variant costs $1.90/$8.00.
Sources
Claude Sonnet 5 — verified August 29, 2026
Kimi K2.7 Code — verified August 29, 2026
Full Claude Sonnet 5 pricing page → · Full Kimi K2.7 Code pricing page →