127 AI models with API pricing under $1 per million output tokens. These ultra-budget options deliver surprising quality — many include vision, reasoning, and function calling capabilities.
| # | Model | Score | Input $/1M | Output $/1M |
|---|---|---|---|---|
| 1 | Gemini 2.5 Flash Lite Preview 09-2025Google | 65 | $0.100 | $0.400 |
| 2 | GPT-5 NanoOpenAI | 64 | $0.050 | $0.400 |
| 3 | Gemini 2.5 Flash LiteGoogle | 64 | $0.100 | $0.400 |
| 4 | Grok 4.1 FastxAI | 64 | $0.200 | $0.500 |
| 5 | Grok 4 FastxAI | 64 | $0.200 | $0.500 |
| 6 | Qwen3.5-FlashAlibaba | 62 | $0.100 | $0.400 |
| 7 | Seed-2.0-MiniByteDance | 61 | $0.100 | $0.400 |
| 8 | Seed 1.6 FlashByteDance | 60 | $0.075 | $0.300 |
| 9 | GPT-4.1 NanoOpenAI | 58 | $0.100 | $0.400 |
| 10 | Gemini 2.0 Flash LiteGoogle | 55 | $0.075 | $0.300 |
| 11 | Qwen Plus 0728 (thinking)Alibaba | 55 | $0.260 | $0.780 |
| 12 | Gemini 2.0 FlashGoogle | 54 | $0.100 | $0.400 |
| 13 | Qwen3 VL 32B InstructAlibaba | 54 | $0.104 | $0.416 |
| 14 | Qwen3 VL 8B InstructAlibaba | 54 | $0.080 | $0.500 |
| 15 | Qwen3 VL 30B A3B InstructAlibaba | 54 | $0.130 | $0.520 |
| 16 | MiMo-V2-FlashXiaomi | 54 | $0.090 | $0.290 |
| 17 | Trinity Miniarcee-ai | 53 | $0.045 | $0.150 |
| 18 | Tongyi DeepResearch 30B A3BAlibaba | 53 | $0.090 | $0.450 |
| 19 | DeepSeek V3.2DeepSeek | 53 | $0.250 | $0.400 |
| 20 | DeepSeek V3.2 ExpDeepSeek | 53 | $0.270 | $0.410 |
| 21 | gpt-oss-safeguard-20bOpenAI | 53 | $0.075 | $0.300 |
| 22 | Mistral Small 3.2 24BMistral AI | 53 | $0.060 | $0.180 |
| 23 | Qwen3 Coder FlashAlibaba | 52 | $0.195 | $0.975 |
| 24 | Llama 4 MaverickMeta | 52 | $0.150 | $0.600 |
| 25 | Qwen Plus 0728Alibaba | 51 | $0.260 | $0.780 |
| 26 | Step 3.5 FlashStepFun | 51 | $0.100 | $0.300 |
| 27 | ERNIE 4.5 VL 28B A3BBaidu | 51 | $0.140 | $0.560 |
| 28 | KAT-Coder-Pro V1Kuaishou | 51 | $0.207 | $0.828 |
| 29 | Nemotron Nano 12B 2 VLNVIDIA | 50 | $0.200 | $0.600 |
| 30 | Llama 4 ScoutMeta | 50 | $0.080 | $0.300 |
| 31 | Qwen3 Coder NextAlibaba | 50 | $0.120 | $0.750 |
| 32 | LongCat Flash ChatMeituan | 50 | $0.200 | $0.800 |
| 33 | Qwen3 30B A3B Instruct 2507Alibaba | 50 | $0.090 | $0.300 |
| 34 | Gemma 3 27BGoogle | 50 | $0.040 | $0.150 |
| 35 | DeepSeek V3.1DeepSeek | 50 | $0.150 | $0.750 |
| 36 | Qwen3 VL 235B A22B InstructAlibaba | 49 | $0.200 | $0.880 |
| 37 | Ministral 3 14B 2512Mistral AI | 49 | $0.200 | $0.200 |
| 38 | Ministral 3 8B 2512Mistral AI | 49 | $0.150 | $0.150 |
| 39 | Olmo 3.1 32B ThinkAllen AI | 49 | $0.150 | $0.500 |
| 40 | Olmo 3 32B ThinkAllen AI | 49 | $0.150 | $0.500 |
At $0.50/1M output tokens, generating a 2,000-word blog post (~2,500 tokens) costs about $0.00125 — roughly 1/10th of a penny. You could generate 800 blog posts for $1. For chatbots, even heavy usage stays under a few dollars per month.
Sub-$1 models have improved dramatically. Many score above 70 on our composite index and include advanced features like vision and reasoning. The quality gap between budget and premium models continues to shrink with each generation.
Budget models are ideal for: high-volume batch processing, simple classification tasks, draft generation with human review, prototype development, and any use case where cost per request matters more than peak quality.
Premium models justify their cost for: complex reasoning tasks, customer-facing applications requiring top accuracy, specialized code generation, and scenarios where errors have high downstream costs.