DeepSeek V4.1 Flash vs Nemotron 3 Ultra: cost per task
DeepSeek V4.1 Flash is cheaper on 14 of 14 tasks and about 16.3x cheaper on average (geometric mean). DeepSeek V4.1 Flash has the higher LMArena Text score (1475 vs 1426).
| Task | DeepSeek V4.1 Flash | Nemotron 3 Ultra | Difference |
|---|---|---|---|
| Answer a chat question | 0.09¢ | 0.33¢ | 3.6x |
| Reply to a customer email | 0.03¢ | 0.86¢ | 28.4x |
| Write a 1,500-word blog post | 0.04¢ | 1.9¢ | 42.0x |
| Write 10 product descriptions | 0.04¢ | 0.75¢ | 18.3x |
| Translate 1,000 words | 0.18¢ | 3.0¢ | 16.8x |
| Proofread one page (500 words) | 0.04¢ | 1.0¢ | 24.1x |
| Summarize a 20-page PDF | 0.12¢ | 1.3¢ | 10.8x |
| Extract data from an invoice | 0.04¢ | 0.66¢ | 14.8x |
| Analyze a 200-row spreadsheet | 0.26¢ | 2.8¢ | 10.7x |
| Answer from 5 retrieved documents (RAG) | 0.08¢ | 0.99¢ | 12.3x |
| Classify 100 support tickets | 0.08¢ | 1.8¢ | 23.0x |
| Review 500 lines of code | 0.26¢ | 2.5¢ | 9.3x |
| Write a SQL query | 0.03¢ | 0.76¢ | 24.1x |
| Run a 50-step coding agent | 2.4¢ | 67¢ | 27.7x |
Specs and list prices
| DeepSeek V4.1 Flash | Nemotron 3 Ultra | |
|---|---|---|
| Input per 1M | $0.072 | $0.60 |
| Output per 1M | $0.14 | $2.40 |
| Cached input per 1M | $0.002 | $0.12 |
| Context | 1M | 262K |
| Providers | 34 | 5 |
| Reasoning | default | default |
| LMArena Text | 1475 | 1426 |
| Released | Sep 10, 2026 | Jun 4, 2026 |
FAQ
Is DeepSeek V4.1 Flash or Nemotron 3 Ultra cheaper?
DeepSeek V4.1 Flash is cheaper on 14 of 14 tasks and about 16.3x cheaper on average (geometric mean). DeepSeek V4.1 Flash has the higher LMArena Text score (1475 vs 1426).
How much does DeepSeek V4.1 Flash cost compared to Nemotron 3 Ultra per token?
DeepSeek V4.1 Flash: $0.072 input and $0.14 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.