Llama 4 Maverick vs DeepSeek V3.2: API pricing comparison
Meta Llama 4 Maverick vs DeepSeek DeepSeek V3.2: token prices, batch and cache rates, Arena ELO, and cost per request for 8 business use cases. Prices in USD per 1 million tokens. Data snapshot: 2026-11-12.
Llama 4 Maverick vs DeepSeek V3.2: the short version
At the prices we quote, DeepSeek V3.2 costs less than Llama 4 Maverick on input and on output: $0.26 against $0.35 per 1M input tokens (26% less), and $0.38 against $1.00 per 1M output tokens (62% less). With the Support Ticket profile (1,500 input and 500 output tokens per request), 1,000 requests cost $1.03 on Llama 4 Maverick and $0.58 on DeepSeek V3.2 at list price, 44% less on DeepSeek V3.2. On Arena ELO, DeepSeek V3.2 scores 1425 and Llama 4 Maverick scores 1287, 138 points apart (scores as of 2026-11-08). The context window is 524K tokens on Llama 4 Maverick and 164K tokens on DeepSeek V3.2. Llama 4 Maverick is open weights (the Llama 4 licence, which restricts some uses); DeepSeek V3.2 is open source (MIT), so it can be self-hosted. Prices as quoted on 2026-11-12. Each price is the list price of one host, the one we quote for that model. A model sold by several hosts has other prices, listed on its model page.
Llama 4 Maverick pricing (Meta)
- Input: $0.35 per 1M tokens
- Output: $1.00 per 1M tokens
- Batch API (input/output): not published
- Prompt-cache read: $0.17
- Context window: 524K
- Arena ELO: 1287 (as of 2026-11-08)
- Full Llama 4 Maverick pricing page
DeepSeek V3.2 pricing (DeepSeek)
- Input: $0.26 per 1M tokens
- Output: $0.38 per 1M tokens
- Batch API (input/output): not published
- Prompt-cache read: $0.13
- Context window: 164K
- Arena ELO: 1425 (as of 2026-11-08)
- Full DeepSeek V3.2 pricing page
Cost per 1,000 requests by business use case
List price vs after discounts (prompt caching on the cached input share, batch API for async workloads, only where the provider publishes those rates).
| Use case | Tokens (in/out) | Llama 4 Maverick list price | Llama 4 Maverick after discounts | DeepSeek V3.2 list price | DeepSeek V3.2 after discounts |
|---|---|---|---|---|---|
| Support Ticket | 1,500 / 500 | $1.03 | $0.86 | $0.58 | $0.46 |
| Knowledge Q&A | 2,000 / 800 | $1.50 | $1.25 | $0.82 | $0.64 |
| Meeting Summary | 10,000 / 1,200 | $4.70 | $4.52 | $3.06 | $2.93 |
| Marketing Content | 2,500 / 1,800 | $2.68 | $2.58 | $1.33 | $1.27 |
| Coding Task | 3,000 / 2,000 | $3.05 | $2.78 | $1.54 | $1.35 |
| Invoice Processing | 1,500 / 600 | $1.13 | $1.04 | $0.62 | $0.56 |
| Call Summary | 2,000 / 700 | $1.40 | $1.36 | $0.79 | $0.76 |
| Agent Workflow | 6,000 / 3,000 | $5.10 | $4.34 | $2.70 | $2.15 |
Common questions about Llama 4 Maverick vs DeepSeek V3.2
- Which is cheaper, Llama 4 Maverick or DeepSeek V3.2?
- At the prices we quote, DeepSeek V3.2 costs less than Llama 4 Maverick on input and on output: $0.26 against $0.35 per 1M input tokens (26% less), and $0.38 against $1.00 per 1M output tokens (62% less). Prices as quoted on 2026-11-12. Each price is the list price of one host, the one we quote for that model. A model sold by several hosts has other prices, listed on its model page.
- What do 1,000 Support Ticket requests cost on Llama 4 Maverick and on DeepSeek V3.2?
- With the Support Ticket profile (1,500 input and 500 output tokens per request), 1,000 requests cost $1.03 on Llama 4 Maverick and $0.58 on DeepSeek V3.2 at list price, 44% less on DeepSeek V3.2. Prices as quoted on 2026-11-12.
- How do Llama 4 Maverick and DeepSeek V3.2 compare on Arena score, context window and licence?
- On Arena ELO, DeepSeek V3.2 scores 1425 and Llama 4 Maverick scores 1287, 138 points apart (scores as of 2026-11-08). The context window is 524K tokens on Llama 4 Maverick and 164K tokens on DeepSeek V3.2. Llama 4 Maverick is open weights (the Llama 4 licence, which restricts some uses); DeepSeek V3.2 is open source (MIT), so it can be self-hosted.
Open this comparison in the interactive compare tool: add up to two more models, copy a shareable link, or export the table as CSV.
These are list prices: real cost also depends on caching, batching, negotiated discounts, marketplace agreements, and existing contracts. Get a FinOps review from OptimNow or compare all models.