Cheapest AI to answer from 5 retrieved documents (Oct 2026)
The cheapest model overall is gpt-oss-20b at 0.01¢ per task. If you want a top-quartile model, the cheapest is GLM 5.3 Flash at 0.06¢.
10 cheapest models
Any model with a price, including small and older ones.
| # | Model | Cost per task | Text score | Tokens in / out | Basis | |
|---|---|---|---|---|---|---|
| 1 |  gpt-oss-20bOpenAI | 0.01¢$0.11 / 1,000 | 1317 | 3,690 / 442 | estimated best value | |
| 2 | S SummarizeScaledown | 0.01¢$0.15 / 1,000 | - | 3,751 / 2,886 | estimated | |
| 3 | S ExtractScaledown | 0.01¢$0.15 / 1,000 | - | 3,751 / 2,886 | estimated | |
| 4 |  Mistral NemoMistral | 0.02¢$0.21 / 1,000 | - | 3,772 / 3,191 | estimated | |
| 5 | I Ling 3.0 FlashInclusionai | 0.03¢$0.26 / 1,000 | - | 3,747 / 2,885 | estimated | |
| 6 | S Llama 3 8B LunarisSao10K | 0.03¢$0.29 / 1,000 | - | 3,751 / 2,886 | estimated | |
| 7 | GPT-5 NanoOpenAI | 0.04¢$0.36 / 1,000 | 1338 | 3,690 / 442 | estimated best value | |
| 8 |  Granite 4.0 MicroIBM | 0.04¢$0.39 / 1,000 | - | 3,721 / 2,878 | estimated | |
| 9 |  Gemma 3 4BGoogle | 0.04¢$0.40 / 1,000 | 1303 | 3,608 / 2,168 | estimated | |
| 10 | gpt-oss-safeguard-20bOpenAI | 0.04¢$0.41 / 1,000 | - | 3,690 / 442 | 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 | 0.06¢$0.57 / 1,000 | 1474 | 3,714 / 1,971 | measured best value | |
| 2 |  DeepSeek V4.1 FlashDeepSeek | 0.08¢$0.81 / 1,000 | 1475 | 3,696 / 3,750 | measured best value | |
| 3 | DeepSeek V4 Pro 0423DeepSeek | 0.23¢$2.34 / 1,000 | 1458 | 3,696 / 3,750 | estimated | |
| 4 |  MiMo-V2.6-ProXiaomi | 0.28¢$2.76 / 1,000 | 1479 | 3,807 / 1,266 | estimated best value | |
| 5 | MiMo-V2.5-ProXiaomi | 0.28¢$2.76 / 1,000 | 1467 | 3,807 / 1,266 | estimated | |
| 6 |  Kimi K2.6Moonshot AI | 0.68¢$6.83 / 1,000 | 1461 | 3,837 / 1,273 | estimated | |
| 7 | GLM 5.2Z.ai | 0.71¢$7.07 / 1,000 | 1476 | 3,714 / 1,971 | estimated | |
| 8 | Gemini 3 Flash PreviewGoogle | 0.83¢$8.31 / 1,000 | 1473 | 3,608 / 2,168 | estimated | |
| 9 | GLM 5.1Z.ai | 0.96¢$9.57 / 1,000 | 1465 | 3,714 / 1,971 | estimated | |
| 10 | GLM 5.3Z.ai | 0.98¢$9.75 / 1,000 | 1478 | 3,714 / 1,971 | estimated |
Cheapest model of each lab
| Model | Cost per task | Score |
|---|---|---|
gpt-oss-20bOpenAI | 0.01¢$0.11 / 1,000 | 1317 |
Mistral NemoMistral | 0.02¢$0.21 / 1,000 | - |
Gemma 3 4BGoogle | 0.04¢$0.40 / 1,000 | 1303 |
 Llama 3.1 8B InstructMeta | 0.04¢$0.42 / 1,000 | 1211 |
 Qwen3.7 FlashAlibaba Qwen | 0.05¢$0.49 / 1,000 | - |
GLM 5.3 FlashZ.ai | 0.06¢$0.57 / 1,000 | 1474 |
DeepSeek V4.1 FlashDeepSeek | 0.08¢$0.81 / 1,000 | 1475 |
MiMo-V2.6-FlashXiaomi | 0.13¢$1.33 / 1,000 | 1451 |
 Claude Haiku 5.5Anthropic | 0.14¢$1.44 / 1,000 | - |
 MiniMax-01MiniMax | 0.21¢$2.14 / 1,000 | - |
Kimi K2.5Moonshot AI | 0.51¢$5.10 / 1,000 | 1450 |
 Grok Build 0.1xAI | 0.71¢$7.14 / 1,000 | - |
FAQ
What is the cheapest AI to answer from 5 retrieved documents?
The cheapest model overall is gpt-oss-20b at 0.01¢ per task. If you want a top-quartile model, the cheapest is GLM 5.3 Flash at 0.06¢.
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.