Llama 4 Maverick vs Nemotron 3 Ultra: cost per task
Llama 4 Maverick is cheaper on 13 of 14 tasks and about 3.9x cheaper on average (geometric mean). Nemotron 3 Ultra has the higher LMArena Text score (1426 vs 1327).
| Task | Llama 4 Maverick | Nemotron 3 Ultra | Difference |
|---|---|---|---|
| Answer a chat question | 0.11¢ | 0.33¢ | 3.1x |
| Reply to a customer email | 0.05¢ | 0.86¢ | 17.0x |
| Write a 1,500-word blog post | 1.0¢ | 1.9¢ | 1.8x |
| Write 10 product descriptions | 1.1¢ | 0.75¢ | 1.4x |
| Translate 1,000 words | 0.11¢ | 3.0¢ | 27.1x |
| Proofread one page (500 words) | 0.13¢ | 1.0¢ | 7.8x |
| Summarize a 20-page PDF | 0.42¢ | 1.3¢ | 3.2x |
| Extract data from an invoice | 0.15¢ | 0.66¢ | 4.3x |
| Analyze a 200-row spreadsheet | 1.2¢ | 2.8¢ | 2.4x |
| Answer from 5 retrieved documents (RAG) | 0.26¢ | 0.99¢ | 3.8x |
| Classify 100 support tickets | 0.58¢ | 1.8¢ | 3.2x |
| Review 500 lines of code | 1.1¢ | 2.5¢ | 2.2x |
| Write a SQL query | 0.17¢ | 0.76¢ | 4.4x |
| Run a 50-step coding agent | 19¢ | 67¢ | 3.6x |
Specs and list prices
| Llama 4 Maverick | Nemotron 3 Ultra | |
|---|---|---|
| Input per 1M | $0.19 | $0.60 |
| Output per 1M | $0.65 | $2.40 |
| Cached input per 1M | $0.05 | $0.12 |
| Context | 1M | 262K |
| Providers | 4 | 5 |
| Reasoning | none | default |
| LMArena Text | 1327 | 1426 |
| Released | Apr 5, 2025 | Jun 4, 2026 |
FAQ
Is Llama 4 Maverick or Nemotron 3 Ultra cheaper?
Llama 4 Maverick is cheaper on 13 of 14 tasks and about 3.9x cheaper on average (geometric mean). Nemotron 3 Ultra has the higher LMArena Text score (1426 vs 1327).
How much does Llama 4 Maverick cost compared to Nemotron 3 Ultra per token?
Llama 4 Maverick: $0.19 input and $0.65 output per 1M tokens. Nemotron 3 Ultra: $0.60 input and $2.40 output. Token prices alone mislead - models use different numbers of tokens for the same task, which is what the task table shows.