DeepSeek V4.1 Flash vs GLM 5.3: cost per task
DeepSeek V4.1 Flash is cheaper on 14 of 14 tasks and about 19.3x cheaper on average (geometric mean). On LMArena Text they are practically tied (1478 vs 1475, within the margin of error).
| Task | DeepSeek V4.1 Flash | GLM 5.3 | Difference |
|---|---|---|---|
| Answer a chat question | 0.09¢ | 0.86¢ | 9.4x |
| Reply to a customer email | 0.03¢ | 0.13¢ | 4.1x |
| Write a 1,500-word blog post | 0.04¢ | 3.3¢ | 75.0x |
| Write 10 product descriptions | 0.04¢ | 3.4¢ | 82.0x |
| Translate 1,000 words | 0.18¢ | 4.9¢ | 27.4x |
| Proofread one page (500 words) | 0.04¢ | 0.39¢ | 9.0x |
| Summarize a 20-page PDF | 0.12¢ | 0.97¢ | 8.0x |
| Extract data from an invoice | 0.04¢ | 0.50¢ | 11.2x |
| Analyze a 200-row spreadsheet | 0.26¢ | 7.7¢ | 29.1x |
| Answer from 5 retrieved documents (RAG) | 0.08¢ | 0.98¢ | 12.1x |
| Classify 100 support tickets | 0.08¢ | 3.8¢ | 47.1x |
| Review 500 lines of code | 0.26¢ | 7.7¢ | 29.1x |
| Write a SQL query | 0.03¢ | 0.77¢ | 24.2x |
| Run a 50-step coding agent | 2.4¢ | 39¢ | 16.0x |
Specs and list prices
| DeepSeek V4.1 Flash | GLM 5.3 | |
|---|---|---|
| Input per 1M | $0.072 | $0.079 |
| Output per 1M | $0.14 | $4.80 |
| Cached input per 1M | $0.002 | $0.078 |
| Context | 1M | 1M |
| Providers | 34 | 46 |
| Reasoning | default | always |
| LMArena Text | 1475 | 1478 |
| Released | Sep 10, 2026 | Aug 18, 2026 |
FAQ
Is DeepSeek V4.1 Flash or GLM 5.3 cheaper?
DeepSeek V4.1 Flash is cheaper on 14 of 14 tasks and about 19.3x cheaper on average (geometric mean). On LMArena Text they are practically tied (1478 vs 1475, within the margin of error).
How much does DeepSeek V4.1 Flash cost compared to GLM 5.3 per token?
DeepSeek V4.1 Flash: $0.072 input and $0.14 output per 1M tokens. GLM 5.3: $0.079 input and $4.80 output. Token prices alone mislead - models use different numbers of tokens for the same task, which is what the task table shows.