Updated 2026-10-11
DeepSeek V4 API Pricing: Pro, Flash and Major Rivals
Current official API rates: Flash — Off-peak: $0.15 input / $0.6 output per 1M tokens (cache-miss input); Peak: $0.3 input / $1.2 output per 1M tokens (cache-miss input). Pro — Off-peak: $0.66 input / $1.98 output per 1M tokens (cache-miss input); Peak: $1.32 input / $3.96 output per 1M tokens (cache-miss input). Peak hours are 09:00–12:00 and 14:00–18:00 Beijing time (01:00–04:00 and 06:00–10:00 UTC), Monday–Friday, excluding Chinese public holidays. All other hours, weekends, and Chinese public holidays are off-peak. Cached input has separate rates; see the complete official table linked from the pricing guide.
Practical verdict
Use DeepSeek V4 as the cost baseline and compare Pro or Flash against the actual workload rather than a vague DeepSeek label. Add paid fallbacks only when measured quality improvements justify the higher cost for a specific request class.
Model snapshot
| Model | Provider | Strengths | Context | Cost signal |
|---|---|---|---|---|
| DeepSeek V4 | DeepSeek | Coding, Multimodal, Long Context | 1M | Provider-specific |
| GPT 5.4 | OpenAI | Reasoning, Tool Calling, Multimodal | 1M | $8.75 / 1M avg tokens |
| Claude Sonnet 4.7 | Anthropic | Coding, Agentic, Long Context | 1M | $9.00 / 1M avg tokens |
| Gemini 3.1 Pro | Reasoning, Multimodal, Long Context | 2M | $7.00 / 1M avg tokens | |
| Qwen 3.5 | Alibaba | Multilingual, Reasoning, Open Source, Cost-Efficiency | 1M | $1.14 / 1M avg tokens |
| MiniMax M2.7 | MiniMax | Agentic, Coding, Long Context, Cost-Efficiency | 205K | $0.75 / 1M avg tokens |
| GLM 5 | Zhipu AI | Coding, Agentic, Multilingual, Cost-Efficiency | 200K | $0.90 / 1M avg tokens |
Cost signals are comparison data used by this site. Verify live provider pricing before production purchasing decisions.
Use-case routing table
| Use case | DeepSeek fit | Alternative fit | Decision note |
|---|---|---|---|
| Default chat and coding API | Best cost baseline | Premium fallback | Start with DeepSeek before paying premium-provider prices on every request. |
| Long-context research | Strong on 1M context | Gemini/Claude strong | Large multimodal inputs can still justify specialized models. |
| Multilingual production | Strong | Qwen/GLM strong | Cost and native-language quality both matter in real deployments. |
| Interactive experience product | Good | MiniMax/Grok strong | Experience quality can justify a different default. |
How to read pricing comparisons in the V4 era
Current official API rates: Flash — Off-peak: $0.15 input / $0.6 output per 1M tokens (cache-miss input); Peak: $0.3 input / $1.2 output per 1M tokens (cache-miss input). Pro — Off-peak: $0.66 input / $1.98 output per 1M tokens (cache-miss input); Peak: $1.32 input / $3.96 output per 1M tokens (cache-miss input). Peak hours are 09:00–12:00 and 14:00–18:00 Beijing time (01:00–04:00 and 06:00–10:00 UTC), Monday–Friday, excluding Chinese public holidays. All other hours, weekends, and Chinese public holidays are off-peak. Cached input has separate rates; see the complete official table linked from the pricing guide.
Why DeepSeek V4 is the baseline
Use the same task and context to compare quality before paying for a more expensive model. Include failed attempts, cached input, output tokens, and human correction in the cost.
Why Flash deserves separate attention
Current official API rates: Flash — Off-peak: $0.15 input / $0.6 output per 1M tokens (cache-miss input); Peak: $0.3 input / $1.2 output per 1M tokens (cache-miss input). Pro — Off-peak: $0.66 input / $1.98 output per 1M tokens (cache-miss input); Peak: $1.32 input / $3.96 output per 1M tokens (cache-miss input). Peak hours are 09:00–12:00 and 14:00–18:00 Beijing time (01:00–04:00 and 06:00–10:00 UTC), Monday–Friday, excluding Chinese public holidays. All other hours, weekends, and Chinese public holidays are off-peak. Cached input has separate rates; see the complete official table linked from the pricing guide.
Where the pricing page fits
The pricing page lists inventory-backed access products. Official API rates, supported tools, quotas, and purchase terms are separate; confirm the selected product before paying.
FAQ
Is DeepSeek V4 the cheapest AI API?
Current official API rates: Flash — Off-peak: $0.15 input / $0.6 output per 1M tokens (cache-miss input); Peak: $0.3 input / $1.2 output per 1M tokens (cache-miss input). Pro — Off-peak: $0.66 input / $1.98 output per 1M tokens (cache-miss input); Peak: $1.32 input / $3.96 output per 1M tokens (cache-miss input). Peak hours are 09:00–12:00 and 14:00–18:00 Beijing time (01:00–04:00 and 06:00–10:00 UTC), Monday–Friday, excluding Chinese public holidays. All other hours, weekends, and Chinese public holidays are off-peak. Cached input has separate rates; see the complete official table linked from the pricing guide.
Is DeepSeek V4 Flash good enough for production?
Yes for many high-volume routes. Flash is excellent for routine coding, chat, retrieval, and tool calls; keep Pro for harder reasoning and review-heavy turns.
What should I compare besides token price?
Compare latency, retries, correctness, context fit, and cost per accepted output.
Why are some compared models not on the pricing page?
Comparison coverage is independent from inventory. Only in-stock Coding Plan products are purchasable.