DeepSeek: R1 vs Meta: Llama 4 Maverick
Meta: Llama 4 Maverick is cheaper than DeepSeek: R1 by about 80% on a blended (75% input / 25% output) basis. DeepSeek: R1 charges $0.280000 / $0.420000 per 1M input/output tokens; Meta: Llama 4 Maverick charges $0.050000 / $0.100000.
| Field | DeepSeek: R1 | Meta: Llama 4 Maverick |
|---|---|---|
| Input $/Mtok | $0.280000 | $0.050000 |
| Output $/Mtok | $0.420000 | $0.100000 |
| Cache read $/Mtok | $0.028000 | – |
| Cache write $/Mtok | – | – |
| Input $/Mchar | 0.091836 | 0.015047 |
| Output $/Mchar | 0.137755 | 0.030093 |
| Context window | 64000 | 1048576 |
| Tokenizer | deepseek | llama3 |
Price history
Same workload, monthly cost
Illustrative monthly cost at 10M tokens/month, split per workload preset. Not a substitute for the calculator, which uses your own volume, cache and batch assumptions.
| Workload | DeepSeek: R1 | Meta: Llama 4 Maverick |
|---|---|---|
| Coding agent | $2.87 | $0.53 |
| RAG chat | $3.01 | $0.58 |
| Summarisation (batch) | $2.94 | $0.55 |
| Classification | $2.83 | $0.51 |
| Chat assistant | $3.36 | $0.70 |
| Content generation | $3.92 | $0.90 |
Cite this page
DeepSeek: R1 vs Meta: Llama 4 Maverick: $0.280000/$0.420000 vs $0.050000/$0.100000 per 1M tokens, as of 2026-10-04 (Token Barometer, https://tokenbarometer.com/compare/deepseek_deepseek-reasoner-vs-meta_llama-4-maverick).
DeepSeek: R1 page | Meta: Llama 4 Maverick page | Open in calculator