Updated October 2026
Cheapest AI models in 2026
The cheapest AI models for everyday chat, ranked by price per million tokens. See input and output prices side by side, plus the cheapest models that still deliver good quality.
Ranking
Best value: cheapest good models
Ranked by blended price per 1M tokens, weighting input 3 to 1 over output to reflect typical chat. Only models scoring 30 or higher on the intelligence index are included, so the cheapest picks here are still good enough to use.
- 1
MiMo-V2.6-FlashXiaomi · $0.14 in / $0.28 out · 1.05M context$0.18
- 2
Claude Haiku 5.5Anthropic · $0.10 in / $0.50 out · 1M context$0.20
- 3
GPT-6 LunaOpenAI · $0.10 in / $0.50 out · 1.05M context$0.20
- 4
GLM 5.3 FlashZ.ai · $0.15 in / $0.50 out · 1.05M context$0.24
- 5
DeepSeek V4 Pro 0423DeepSeek · $0.21 in / $0.42 out · 1.05M context$0.26 - 6
DeepSeek V4 Flash Vision ExpDeepSeek · $0.22 in / $0.65 out · 1.05M context$0.32 - 7
DeepSeek V4 Flash 0731DeepSeek · $0.00 in / $1.28 out · 1.05M context$0.32 - 8
GPT-5.6 LunaOpenAI · $0.20 in / $1.20 out · 1.05M context$0.45
- 9
DeepSeek V4.1 FlashDeepSeek · $0.30 in / $1.20 out · 1.05M context$0.52 - 10
MiMo-V2.6-ProXiaomi · $0.44 in / $0.87 out · 1.05M context$0.54
- 11
Qwen3.8 27BQwen · $0.43 in / $2.55 out · 1M context$0.96 - 12
DeepSeek V4 Pro 0813DeepSeek · $0.66 in / $1.98 out · 1.05M context$0.99 - 13
Gemini 3.8 FlashGoogle · $0.75 in / $3.75 out · 1.05M context$1.50
- 14
Gemini 3.7 FlashGoogle · $0.75 in / $3.75 out · 1.05M context$1.50
- 15
Gemini 3.6 FlashGoogle · $0.75 in / $3.75 out · 1.05M context$1.50
- 16
GLM 5.3Z.ai · $0.04 in / $7 out · 1.05M context$1.78
- 17
Muse Spark 1.3Meta · $1.25 in / $4.25 out · 1.05M context$2
- 18
Muse Spark 1.2Meta · $1.25 in / $4.25 out · 1.05M context$2
- 19
Muse Spark 1.1Meta · $1.25 in / $4.25 out · 1.05M context$2
- 20
GLM 5.2Z.ai · $0.06 in / $8 out · 1.05M context$2.05
No quality filter
Absolute cheapest, any quality
Every model ranked purely by price, including basic ones that score below 30 on the intelligence index. Fine for simple, high-volume jobs like tagging or sorting. Check the quality column before you commit.
- 1
Mistral NemoMistral · $0.03 in / $0.03 out · 131K context · Quality: Unrated$0.03 - 2
Ling 3.0 Flash VLinclusionAI · $0.02 in / $0.06 out · 262K context · Quality: Basic$0.03
- 3
Ling 3.0 FlashinclusionAI · $0.02 in / $0.06 out · 262K context · Quality: Basic$0.03
- 4
gpt-oss-20bOpenAI · $0.02 in / $0.09 out · 131K context · Quality: Basic$0.04
- 5
Granite 4.0 MicroIBM · $0.02 in / $0.11 out · 131K context · Quality: Basic$0.04 - 6
Qwen3.7 FlashQwen · $0.03 in / $0.13 out · 1M context · Quality: Unrated$0.06 - 7
Llama 3.1 8B InstructMeta · $0.05 in / $0.08 out · 131K context · Quality: Basic$0.06
- 8
Mistral Small 3Mistral · $0.05 in / $0.08 out · 33K context · Quality: Basic$0.06 - 9
Nova Micro 1.0Amazon · $0.04 in / $0.14 out · 128K context · Quality: Basic$0.06
- 10
Ling 3.0 Flash FininclusionAI · $0.04 in / $0.12 out · 262K context · Quality: Basic$0.06
- 11
Gemma 3 4BGoogle · $0.05 in / $0.10 out · 131K context · Quality: Basic$0.06
- 12
Mercury 2.5Inception · $0.04 in / $0.15 out · 260K context · Quality: Basic$0.07
- 13
Nemotron 3.5 LightningNVIDIA · $0.05 in / $0.13 out · 262K context · Quality: Basic$0.07
- 14
gpt-oss-120bOpenAI · $0.04 in / $0.17 out · 131K context · Quality: Basic$0.07
- 15
Llama 3.2 1B InstructMeta · $0.03 in / $0.20 out · 60K context · Quality: Basic$0.07
- 16
Gemma 3 12BGoogle · $0.05 in / $0.15 out · 131K context · Quality: Basic$0.08
- 17
Solar Mini 4Upstage · $0.05 in / $0.20 out · 524K context · Quality: Basic$0.09
- 18
Phi 4Microsoft · $0.07 in / $0.14 out · 16K context · Quality: Basic$0.09
- 19
Ministral 3 3B 2512Mistral · $0.10 in / $0.10 out · 131K context · Quality: Basic$0.10 - 20
Nemotron 3 Nano 30B A3BNVIDIA · $0.06 in / $0.24 out · 262K context · Quality: Basic$0.11
Head to head
How the top models compare
Detailed comparisons between the models on this list.
DeepSeek V4.1 Flash vs GLM 5.3 Flash
DeepSeek vs Z.ai
See comparisonMiMo-V2.6-Flash vs DeepSeek V4.1 Flash
Xiaomi vs DeepSeek
See comparisonMiMo-V2.6-Flash vs GLM 5.3 Flash
Xiaomi vs Z.ai
See comparisonGPT-6 Luna vs DeepSeek V4.1 Flash
OpenAI vs DeepSeek
See comparisonGPT-6 Luna vs GLM 5.3 Flash
OpenAI vs Z.ai
See comparisonDeepSeek V4.1 Flash vs DeepSeek V4 Flash 0731
DeepSeek vs DeepSeek
See comparisonMore rankings
Best models for other jobs
The leader changes with the job. See how these same models rank for something else.
Smartest AI models
The models with the highest general intelligence scores, covering reasoning, math, science, and knowledge.
- 1
Claude Opus 5.557.6
- 2
Claude Sonnet 5.556
- 3
Claude Fable 5.153.4
Best AI models for coding
The strongest models for writing, refactoring, and debugging code, ranked by independent coding benchmarks.
- 1
Claude Fable 5.181.6
- 2
Claude Opus 578
- 3
GPT-5.6 Sol77.4
Best long-context AI models
Models that can read entire codebases, books, or long document sets in a single prompt.
- 1
Grok 4.202M tokens
- 2
Llama 4 Scout1.31M tokens
- 3
GPT-6 Astra1.05M tokens
FAQ
Frequently asked questions
What is the cheapest AI model?
MiMo-V2.6-Flash from Xiaomi ranks first with blended price per 1M tokens of $0.18. Claude Haiku 5.5 is second at $0.20.
How are these models ranked?
Ranked by blended price per 1M tokens, weighting input 3 to 1 over output to reflect typical chat.
Can I use more than one of these models?
Yes. Shortcut Chat sends one prompt to several models at once, so you can run the top models on this list side by side and keep the best answer.
Use the top models together.
Send one prompt to multiple models on this list at once, see the answers side by side, and take the best part from each.