Cheapest LLM APIs in 2026
A data-driven ranking of all 19 AI models by estimated monthly cost. Based on 5M input + 2M output tokens — a typical monthly volume for a solo developer or small team. Updated 2026-07-14.
Full Ranking — Cheapest to Most Expensive
| # | Model | Input $/M | Output $/M | Est. Monthly |
|---|---|---|---|---|
| 1 | Command A+ | $0.00 | $0.00 | $0.00 |
| 2 | Mistral Small 3.2 | $0.07 | $0.20 | $0.78 |
| 3 | Llama 4 Scout | $0.10 | $0.30 | $1.10 |
| 4 | DeepSeek V4 Flash | $0.14 | $0.28 | $1.26 |
| 5 | Gemini 2.5 Flash-Lite | $0.10 | $0.40 | $1.30 |
| 6 | GPT-5.4 nano | $0.20 | $1.25 | $3.50 |
| 7 | MiniMax M3 | $0.30 | $1.20 | $3.90 |
| 8 | DeepSeek V4 Pro | $0.43 | $0.87 | $3.92 |
| 9 | Qwen 3 Coder | $0.30 | $1.50 | $4.50 |
| 10 | Mistral Large 3 | $0.50 | $1.50 | $5.50 |
| 11 | Llama 4 Maverick | $0.80 | $2.00 | $8.00 |
| 12 | Grok 4.3 | $1.25 | $2.50 | $11.25 |
| 13 | GPT-5.4 mini | $0.75 | $4.50 | $12.75 |
| 14 | Kimi K2.6 | $0.95 | $4.00 | $12.75 |
| 15 | Claude Haiku 4.5 | $1.00 | $5.00 | $15.00 |
| 16 | GLM-5.2 | $1.40 | $4.40 | $15.80 |
| 17 | GPT-5.6 Luna | $1.00 | $6.00 | $17.00 |
| 18 | Grok 4.5 | $2.00 | $6.00 | $22.00 |
| 19 | Gemini 3.5 Flash | $1.50 | $9.00 | $25.50 |
| 20 | Qwen 3.7 Max | $2.50 | $7.50 | $27.50 |
| 21 | Claude Sonnet 5 | $2.00 | $10.00 | $30.00 |
| 22 | Gemini 3.1 Pro | $2.00 | $12.00 | $34.00 |
| 23 | Nova Premier | $2.50 | $12.50 | $37.50 |
| 24 | GPT-5.6 Terra | $2.50 | $15.00 | $42.50 |
| 25 | GPT-5.4 | $2.50 | $15.00 | $42.50 |
| 26 | GPT-5.3 Codex | $2.50 | $15.00 | $42.50 |
| 27 | Claude Sonnet 4.6 | $3.00 | $15.00 | $45.00 |
| 28 | Claude Opus 4.8 | $5.00 | $25.00 | $75.00 |
| 29 | GPT-5.6 Sol | $5.00 | $30.00 | $85.00 |
| 30 | Claude Fable 5 | $10.00 | $50.00 | $150.00 |
Estimated cost: 5,000,000 input + 2,000,000 output tokens/month, no caching, no batch. Your actual cost depends on your exact usage. Use the calculator for custom estimates →
Key Insights
115×
Price gap between cheapest (Gemini Flash Lite at $1.30/mo) and most expensive (Claude Fable 5 at $150/mo) for the same token volume.
$0.10/M
Gemini Flash Lite has the lowest input price of any model. At 10M input tokens/month, that's just $1.00.
Top 5
The top 5 cheapest models all come from Google (2), DeepSeek (2), and Alibaba (1). No Western provider cracks the top 5.
80.5
MiniMax M3's SWE-bench score at $0.30/M input — matching or beating models that cost 10× more.
Top 3 Cheapest Models — Detailed
#1 — Command A+
Cohere
#2 — Mistral Small 3.2
Mistral
#3 — Llama 4 Scout
Meta
Compare the Cheapest Models
📊 Find the cheapest model for your exact usage
Token volumes, caching, and input/output ratios change the ranking. Enter your numbers to see which model is truly cheapest for you.
Try the Calculator →