Cheapest AI to write a customer email reply (Oct 2026)
The cheapest model overall is Summarize at <0.01¢ per task. If you want a top-quartile model, the cheapest is GLM 5.3 Flash at <0.01¢.
10 cheapest models
Any model with a price, including small and older ones.
| # | Model | Cost per task | Text score | Tokens in / out | Basis | |
|---|---|---|---|---|---|---|
| 1 | S SummarizeScaledown | <0.01¢$0.01 / 1,000 | - | 326 / 685 | estimated | |
| 2 | S ExtractScaledown | <0.01¢$0.01 / 1,000 | - | 326 / 685 | estimated | |
| 3 |  gpt-oss-20bOpenAI | <0.01¢$0.02 / 1,000 | 1317 | 258 / 126 | estimated best value | |
| 4 | S Llama 3 8B LunarisSao10K | <0.01¢$0.05 / 1,000 | - | 326 / 685 | estimated | |
| 5 | I Ling 3.0 FlashInclusionai | <0.01¢$0.05 / 1,000 | - | 325 / 684 | estimated | |
| 6 | gpt-oss-safeguard-20bOpenAI | <0.01¢$0.06 / 1,000 | - | 258 / 126 | estimated | |
| 7 | GPT-5 NanoOpenAI | <0.01¢$0.06 / 1,000 | 1338 | 258 / 126 | estimated best value | |
| 8 | GPT-4.1 NanoOpenAI | <0.01¢$0.06 / 1,000 | 1322 | 258 / 94 | estimated | |
| 9 |  GLM 5.3 FlashZ.ai | <0.01¢$0.06 / 1,000 | 1474 | 265 / 258 | measured best value | |
| 10 |  Gemma 3 4BGoogle | <0.01¢$0.07 / 1,000 | 1303 | 271 / 552 | estimated |
Cheapest high-quality models
Only models in the top 25% of LMArena Text scores (≥ 1458).
| # | Model | Cost per task | Text score | Tokens in / out | Basis | |
|---|---|---|---|---|---|---|
| 1 | GLM 5.3 FlashZ.ai | <0.01¢$0.06 / 1,000 | 1474 | 265 / 258 | measured best value | |
| 2 |  DeepSeek V4.1 FlashDeepSeek | 0.03¢$0.30 / 1,000 | 1475 | 279 / 1,974 | measured best value | |
| 3 |  MiMo-V2.6-ProXiaomi | 0.06¢$0.57 / 1,000 | 1479 | 420 / 449 | estimated best value | |
| 4 | MiMo-V2.5-ProXiaomi | 0.06¢$0.57 / 1,000 | 1467 | 420 / 449 | estimated | |
| 5 | GLM 5.2Z.ai | 0.08¢$0.79 / 1,000 | 1476 | 265 / 258 | estimated | |
| 6 | DeepSeek V4 Pro 0423DeepSeek | 0.09¢$0.88 / 1,000 | 1458 | 279 / 1,974 | estimated | |
| 7 | GLM 5.1Z.ai | 0.10¢$1.04 / 1,000 | 1465 | 265 / 258 | estimated | |
| 8 | GLM 5.3Z.ai | 0.13¢$1.26 / 1,000 | 1478 | 265 / 258 | estimated | |
| 9 | GPT-6 SolOpenAI | 0.15¢$1.46 / 1,000 | 1459 | 258 / 94 | estimated | |
| 10 | GPT-5.6 TerraOpenAI | 0.16¢$1.64 / 1,000 | 1466 | 258 / 94 | estimated |
Cheapest model of each lab
| Model | Cost per task | Score |
|---|---|---|
gpt-oss-20bOpenAI | <0.01¢$0.02 / 1,000 | 1317 |
GLM 5.3 FlashZ.ai | <0.01¢$0.06 / 1,000 | 1474 |
Gemma 3 4BGoogle | <0.01¢$0.07 / 1,000 | 1303 |
 Llama 3.1 8B InstructMeta | <0.01¢$0.07 / 1,000 | 1211 |
 Qwen3.7 FlashAlibaba Qwen | <0.01¢$0.10 / 1,000 | - |
 Mistral NemoMistral | 0.01¢$0.11 / 1,000 | - |
MiMo-V2.6-FlashXiaomi | 0.02¢$0.24 / 1,000 | 1451 |
DeepSeek V4.1 FlashDeepSeek | 0.03¢$0.30 / 1,000 | 1475 |
 Claude Haiku 5.5Anthropic | 0.04¢$0.35 / 1,000 | - |
 MiniMax M2.5MiniMax | 0.05¢$0.54 / 1,000 | 1391 |
 Kimi K2 0711Moonshot AI | 0.13¢$1.28 / 1,000 | 1418 |
 Grok Build 0.1xAI | 0.21¢$2.15 / 1,000 | - |
FAQ
What is the cheapest AI to write a customer email reply?
The cheapest model overall is Summarize at <0.01¢ per task. If you want a top-quartile model, the cheapest is GLM 5.3 Flash at <0.01¢.
Is the cheapest model good enough?
Often not for hard tasks. That is why this page also lists the cheapest models in the top quarter of LMArena scores and the cheapest model of every major lab.