Kimi vs DeepSeek pricing and capability comparison: context, API cost, code workflows, and model choices
Editorial standards and source policy: Editorial standards, Team. Content links to primary sources; see Methodology.
As of July 2, 2026, DeepSeek-V4-Flash and DeepSeek-V4-Pro are the API cost and 1M-context baselines; Kimi K2.x models sit at 262,144-token context and higher API prices but connect to Kimi Code, K2.7 Code, multimodal inputs, and coding workflows.
| Model | Context | Max output | Input cache hit | Input cache miss | Output | Extra signal |
|---|---|---|---|---|---|---|
deepseek-v4-flash |
1M | 384K | $0.0028 / 1M tokens | $0.14 / 1M tokens | $0.28 / 1M tokens | concurrency 2500 |
deepseek-v4-pro |
1M | 384K | $0.003625 / 1M tokens | $0.435 / 1M tokens | $0.87 / 1M tokens | concurrency 500 |
kimi-k2.7-code |
262,144 tokens | not separately listed | $0.19 / 1M tokens | $0.95 / 1M tokens | $4.00 / 1M tokens | Kimi coding model |
kimi-k2.7-code-highspeed |
262,144 tokens | not separately listed | $0.38 / 1M tokens | $1.90 / 1M tokens | $8.00 / 1M tokens | about 180 tokens/s, up to about 260 tokens/s short context |
kimi-k2.6 |
262,144 tokens | not separately listed | $0.16 / 1M tokens | $0.95 / 1M tokens | $4.00 / 1M tokens | general multimodal / agent |
kimi-k2.5 |
262,144 tokens | not separately listed | $0.10 / 1M tokens | $0.60 / 1M tokens | $3.00 / 1M tokens | lower-cost Kimi K2.x |
moonshot-v1-128k |
131,072 tokens | not listed here | n/a | $2.00 / 1M input tokens | $5.00 / 1M output tokens | V1 long-context reference |
| 50k input + 10k output sample | Cache miss cost | Cache hit cost | Read this as |
|---|---|---|---|
deepseek-v4-flash |
$0.0098 | $0.00294 | lowest API cost baseline here |
deepseek-v4-pro |
$0.03045 | $0.00888 | stronger DeepSeek tier, still cheaper than Kimi K2.x in this sample |
kimi-k2.5 |
$0.0600 | $0.0350 | cheapest Kimi option in this set |
kimi-k2.6 |
$0.0875 | $0.0480 | general Kimi multimodal/agent option |
kimi-k2.7-code |
$0.0875 | $0.0495 | Kimi coding model baseline |
kimi-k2.7-code-highspeed |
$0.1750 | $0.0990 | higher price for interactive speed |
For API batch work, long-document summarization, structured extraction, and cost-sensitive workloads, put DeepSeek-V4-Flash in the first test group. For terminal/IDE coding-agent work, Kimi Code plus kimi-k2.7-code deserves a separate pilot because workflow value is not captured by API token price alone.
DeepSeek says deepseek-chat and deepseek-reasoner are deprecated on July 24, 2026 15:59 UTC and map to DeepSeek-V4-Flash non-thinking and thinking modes. Kimi says older K2 preview/thinking model names stopped maintenance on May 25, 2026; kimi-latest stopped maintenance on January 28, 2026; kimi-thinking-preview stopped maintenance on November 11, 2025.
Official sources: Kimi model list, Kimi K2.7 Code pricing, Kimi K2.6 pricing, Kimi K2.5 pricing, Moonshot V1 pricing, Kimi Code GitHub, DeepSeek models pricing, DeepSeek API docs.
Related reading
- Top China-Built AI Models to Watch in 2026: DeepSeek, Qwen, Kimi & More
- Kimi Pricing and API Access for Builders
- Kimi Code CLI and IDE Evaluation for Builders
- How to Use the DeepSeek API for Builder Workflows
RadarAI helps builders track AI updates, compare source-backed signals, and decide which changes are worth acting on.