GLM 5.3 vs Kimi K3: cost per task
GLM 5.3 is cheaper on 14 of 14 tasks and about 3.7x cheaper on average (geometric mean). On LMArena Text they are practically tied (1487 vs 1478, within the margin of error).
| Task | GLM 5.3 | Kimi K3 | Difference |
|---|---|---|---|
| Answer a chat question | 0.86¢ | 1.6¢ | 1.9x |
| Reply to a customer email | 0.13¢ | 4.2¢ | 33.5x |
| Write a 1,500-word blog post | 3.3¢ | 9.2¢ | 2.8x |
| Write 10 product descriptions | 3.4¢ | 3.7¢ | 1.1x |
| Translate 1,000 words | 4.9¢ | 14¢ | 3.0x |
| Proofread one page (500 words) | 0.39¢ | 4.9¢ | 12.7x |
| Summarize a 20-page PDF | 0.97¢ | 3.0¢ | 3.1x |
| Extract data from an invoice | 0.50¢ | 3.2¢ | 6.3x |
| Analyze a 200-row spreadsheet | 7.7¢ | 13¢ | 1.7x |
| Answer from 5 retrieved documents (RAG) | 0.98¢ | 3.9¢ | 4.0x |
| Classify 100 support tickets | 3.8¢ | 8.8¢ | 2.3x |
| Review 500 lines of code | 7.7¢ | 11¢ | 1.4x |
| Write a SQL query | 0.77¢ | 3.8¢ | 4.9x |
| Run a 50-step coding agent | 39¢ | $2.24 | 5.8x |
Specs and list prices
| GLM 5.3 | Kimi K3 | |
|---|---|---|
| Input per 1M | $0.079 | $0.30 |
| Output per 1M | $4.80 | $11.94 |
| Cached input per 1M | $0.078 | $0.21 |
| Context | 1M | 1M |
| Providers | 46 | 26 |
| Reasoning | always | default |
| LMArena Text | 1478 | 1487 |
| Released | Aug 18, 2026 | Jul 16, 2026 |
FAQ
Is GLM 5.3 or Kimi K3 cheaper?
GLM 5.3 is cheaper on 14 of 14 tasks and about 3.7x cheaper on average (geometric mean). On LMArena Text they are practically tied (1487 vs 1478, within the margin of error).
How much does GLM 5.3 cost compared to Kimi K3 per token?
GLM 5.3: $0.079 input and $4.80 output per 1M tokens. Kimi K3: $0.30 input and $11.94 output. Token prices alone mislead - models use different numbers of tokens for the same task, which is what the task table shows.