LLMs under $20 per million tokens
Every model priced below $20 per 1M input tokens. Ranked by benchmark score within budget. The "I care about quality" tier.
Ranked by quality within budget
Input under $20/M, highest benchmark score first.
| # | Model | In $/1M | Out $/1M | Type |
|---|---|---|---|---|
| 1 |  GPT-5 Chat | $1.25 | $10.00 | Closed |
| 2 |  Qwen3.5 397B A17B | $0.39 | $2.34 | OSS |
| 3 |  DeepSeek V3.2 Speciale | $0.40 | $1.20 | OSS |
| 4 |  Gemini 2.5 Pro Preview 05-06 | $1.25 | $10.00 | Closed |
| 5 |  Step 3.5 Flash | $0.10 | $0.30 | OSS |
| 6 |  MiMo-V2-Flash | $0.09 | $0.29 | OSS |
| 7 | GPT-5.1-Codex-Max | $1.25 | $10.00 | Closed |
| 8 | o4 Mini High | $1.10 | $4.40 | Closed |
| 9 | Qwen3.6 Plus | $0.33 | $1.95 | OSS |
| 10 | GPT-5.2-Codex | $1.75 | $14.00 | Closed |
| 11 |  GLM 5.1 | $0.95 | $3.15 | OSS |
| 12 |  Palmyra X5 | $0.60 | $6.00 | Closed |
| 13 |  Grok 3 Beta | $3.00 | $15.00 | Closed |
| 14 |  MiniMax M2 | $0.26 | $1.00 | OSS |
| 15 | GLM 4.5 | $0.60 | $2.20 | OSS |
| 16 | gpt-oss-120b (free) | $0.00 | $0.00 | OSS |
| 17 | GPT-5.1-Codex | $1.25 | $10.00 | Closed |
| 18 | gpt-oss-20b | $0.03 | $0.14 | OSS |
| 19 | gpt-oss-20b (free) | $0.00 | $0.00 | OSS |
| 20 | Qwen3 30B A3B Thinking 2507 | $0.08 | $0.40 | OSS |
| 21 | GPT-4 Turbo (older v1106) | $10.00 | $30.00 | Closed |
| 22 | Grok 3 Mini Beta | $0.30 | $0.50 | Closed |
| 23 | Gemini 2.0 Flash Lite | $0.07 | $0.30 | Closed |
| 24 | MiniMax M2.7 | $0.30 | $1.20 | OSS |
| 25 | Gemma 4 31B | $0.13 | $0.38 | OSS |
| 26 | Qwen3 Next 80B A3B Thinking | $0.10 | $0.78 | OSS |
| 27 | Gemini 3.1 Pro Preview | $2.00 | $12.00 | Closed |
| 28 | GPT-5.1-Codex-Mini | $0.25 | $2.00 | Closed |
| 29 | o3 Mini High | $1.10 | $4.40 | Closed |
| 30 |  LongCat Flash Chat | $0.20 | $0.80 | OSS |
| 31 | Gemini 2.5 Flash Lite | $0.10 | $0.40 | Closed |
| 32 | DeepSeek V3 | $0.32 | $0.89 | OSS |
| 33 | GPT-5.4 | $2.50 | $15.00 | Closed |
| 34 | Qwen3 Max | $0.78 | $3.90 | OSS |
| 35 | Qwen3 32B | $0.08 | $0.24 | OSS |
| 36 | MiMo-V2-Pro | $1.00 | $3.00 | Closed |
| 37 | R1 0528 | $0.50 | $2.15 | OSS |
| 38 |  Mixtral 8x7B Instruct | $0.54 | $0.54 | OSS |
| 39 | GLM 5 | $0.72 | $2.30 | OSS |
| 40 |  Claude Opus 4.6 | $5.00 | $25.00 | Closed |
Top 3 best-quality sub-$20 models
GPT-5 Chat sits in the under-$20 tier with a top-shelf quality score. Appropriate when reasoning or reliability matters and you do not need the absolute frontier.
Qwen3.5 397B A17B sits in the under-$20 tier with a top-shelf quality score. Appropriate when reasoning or reliability matters and you do not need the absolute frontier.
DeepSeek V3.2 Speciale sits in the under-$20 tier with a top-shelf quality score. Appropriate when reasoning or reliability matters and you do not need the absolute frontier.
The price gap · cheapest vs most expensive
At the top of this tier, pricing approaches frontier. The delta to true frontier (Opus, GPT-5) is 2x to 5x for roughly 5 to 15 percent quality gain on most tasks.