Updated 12 Oct 2026 · 286 models checked
The cheapest AI models right now
The cheapest paid model is gpt-oss-20b at per 1M tokens. With a 1M-token context window, Qwen3.7 Flash costs $0.03 in / $0.13 out. 10 models are listed at no per-token charge.
$0.02
derived We computed
- Source
- Our catalogue
- Retrieved
- 10 Oct 2026
- Note
- Lowest standard-tier input price in our catalogue.
Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
Cite
Cite this page
ModelsAtlas. “The cheapest AI models right now”. Retrieved 12 Oct 2026. https://modelsatlas.com/cheapest
Free AI models
All 8 free-tier offers →| # | Model | Input / 1M | Output / 1M | Context | Listed as | |
|---|---|---|---|---|---|---|
| 1 | Lyria 3 Clip PreviewGoogle | Free | Free | 1.05M | 1M context | Alternatives |
| 2 | Lyria 3 Pro PreviewGoogle | Free | Free | 1.05M | 1M context | Alternatives |
| 3 | Nemotron 3 Ultra (free)NVIDIA | Free | Free | 1M | 1M context | Alternatives |
| 4 | Nemotron 3.5 Lightning (free)NVIDIA | Free | Free | 1M | 1M context | Alternatives |
| 5 | Gemma 4 26B A4BGoogle | Free | Free | 262K | Alternatives | |
| 6 | Gemma 4 31BGoogle | Free | Free | 262K | Alternatives | |
| 7 | Nemotron 3 SuperNVIDIA | Free | Free | 262K | Alternatives | |
| 8 | Nemotron 3 Nano Omni (free)NVIDIA | Free | Free | 256K | ||
| 9 | North Mini Code (free)Cohere | Free | Free | 256K | ||
| 10 | Nemotron 3.5 Content SafetyNVIDIA | Free | Free | 131K | Alternatives |
Under $1 per 1M tokens
Full price list →| # | Model | Input / 1M | Output / 1M | Context | Listed as | |
|---|---|---|---|---|---|---|
| 1 | gpt-oss-20bOpenAI | $0.02 | $0.09 | 131K | Alternatives | |
| 2 | Llama 3.2 1B InstructMeta | $0.03 | $0.20 | 60K | Alternatives | |
| 3 | Mistral NemoMistral AI | $0.03 | $0.03 | 131K | Alternatives | |
| 4 | Qwen3.7 FlashQwen (Alibaba) | $0.03 | $0.13 | 1M | 1M context | Alternatives |
| 5 | gpt-oss-120bOpenAI | $0.03 | $0.15 | 131K | Alternatives | |
| 6 | Command R7B (12-2024)Cohere | $0.04 | $0.15 | 128K | Alternatives | |
| 7 | Llama 3.1 8B InstructMeta | $0.05 | $0.08 | 131K | Alternatives | |
| 8 | Mistral Small 3Mistral AI | $0.05 | $0.08 | 33K | Alternatives | |
| 9 | Gemma 3 4BGoogle | $0.05 | $0.10 | 131K | Alternatives | |
| 10 | Gemma 3 12BGoogle | $0.05 | $0.15 | 131K | Alternatives | |
| 11 | Llama 3.2 3B InstructMeta | $0.05 | $0.33 | 131K | Alternatives | |
| 12 | Nemotron 3 Nano 30B A3BNVIDIA | $0.06 | $0.24 | 262K | Alternatives | |
| 13 | Qwen3.5-FlashQwen (Alibaba) | $0.07 | $0.26 | 1M | 1M context | Alternatives |
| 14 | Gemma 4 26B A4BGoogle | $0.07 | $0.23 | 262K | Alternatives | |
| 15 | Phi 4Microsoft | $0.07 | $0.14 | 16K | Alternatives |
$1–$5 per 1M tokens
Full price list →| # | Model | Input / 1M | Output / 1M | Context | Listed as | |
|---|---|---|---|---|---|---|
| 1 | Grok Build 0.1xAI | $1.00 | $2.00 | 256K | Alternatives | |
| 2 | Claude Haiku 4.5Anthropic | $1.00 | $5.00 | 200K | Alternatives | |
| 3 | o3 Mini HighOpenAI | $1.10 | $4.40 | 200K | Alternatives | |
| 4 | o4 MiniOpenAI | $1.10 | $4.40 | 200K | Alternatives | |
| 5 | o4 Mini HighOpenAI | $1.10 | $4.40 | 200K | Alternatives | |
| 6 | Grok 4.20xAI | $1.25 | $2.50 | 2M | 1M context | Alternatives |
| 7 | Grok 4.20 Multi-AgentxAI | $1.25 | $2.50 | 2M | 1M context | Alternatives |
| 8 | Grok 4.3xAI | $1.25 | $2.50 | 1M | 1M context | Alternatives |
| 9 | Muse Spark 1.1Meta | $1.25 | $4.25 | 1.05M | 1M context | Alternatives |
| 10 | Muse Spark 1.2Meta | $1.25 | $4.25 | 1.05M | 1M context | Alternatives |
| 11 | Muse Spark 1.3Meta | $1.25 | $4.25 | 1.05M | 1M context | Alternatives |
| 12 | Gemini 2.5 Pro Preview 06-05Google | $1.25 | $10.00 | 1.05M | 1M context | Alternatives |
| 13 | GPT-5OpenAI | $1.25 | $10.00 | 400K | Alternatives | |
| 14 | GPT-5.1OpenAI | $1.25 | $10.00 | 400K | Alternatives | |
| 15 | GPT-5.1-CodexOpenAI | $1.25 | $10.00 | 400K | Alternatives |
Cheapest with the capability you need
1M context
Qwen3.7 Flash
Qwen (Alibaba)
$0.03
input / 1M · $0.13 output
Open weights
Qwen2.5 7B Instruct
Qwen (Alibaba) · open weights
$0.10
input / 1M · $0.20 output
Extended thinking
Qwen3 235B A22B Instruct 2507
Qwen (Alibaba)
$0.09
input / 1M · $0.55 output
Cheapest paid model whose record lists the capability; a model with nothing recorded is not counted.
Questions
What is the cheapest AI model API?
gpt-oss-20b at $0.02 per 1M input tokens and $0.09 per 1M output is the cheapest paid model we list. 10 models are listed at no per-token charge.
Which models are free?
10 models are listed at no per-token charge; the one with the widest context window is Lyria 3 Clip Preview at 1.05M tokens. A zero price is what our catalogue records — limits and logging terms are the provider's, so read them before you rely on it.
Is the cheapest model good enough?
We do not benchmark quality, so price alone cannot tell you. Check the model's own page for its context window and listed capabilities, and compare it with the next price band before you commit.
How can I pay less for the same model?
Use the batch tier for work that can wait. 52 models list a batch offer; where the same model lists both, batch is a median 50% of standard.
Keep going in your AI
Opens with this page as the source. We add nothing else to the prompt.
- What is the cheapest AI model API?
- Which models are free?
- Is the cheapest model good enough?