Cheapest coding LLMs
The cheapest models that hold up on SWE-bench, HumanEval, MBPP, LiveCodeBench. Ranked by price per 1M input tokens.
Ranked by input price
Models with meaningful coding benchmark scores, cheapest first.
| # | Model | In $/1M | Out $/1M | Type |
|---|---|---|---|---|
| 1 |  gpt-oss-120b (free) | $0.00 | $0.00 | OSS |
| 2 | gpt-oss-20b (free) | $0.00 | $0.00 | OSS |
| 3 | gpt-oss-120b | $0.04 | $0.19 | OSS |
| 4 | GPT-5 Nano | $0.05 | $0.40 | Closed |
| 5 |  Qwen3 8B | $0.05 | $0.40 | OSS |
| 6 |  ERNIE 4.5 21B A3B Thinking | $0.07 | $0.28 | OSS |
| 7 | Qwen3 235B A22B Instruct 2507 | $0.07 | $0.10 | OSS |
| 8 |  Gemma 3 27B | $0.08 | $0.16 | OSS |
| 9 |  Llama 4 Scout | $0.08 | $0.30 | OSS |
| 10 | Qwen3 30B A3B Thinking 2507 | $0.08 | $0.40 | OSS |
| 11 | Qwen3 32B | $0.08 | $0.24 | OSS |
| 12 |  MiMo-V2-Flash | $0.09 | $0.29 | OSS |
| 13 | Qwen3 30B A3B Instruct 2507 | $0.09 | $0.30 | OSS |
| 14 | Qwen3 Next 80B A3B Instruct | $0.09 | $1.10 | OSS |
| 15 | Qwen3 Next 80B A3B Thinking | $0.10 | $0.78 | OSS |
| 16 |  Step 3.5 Flash | $0.10 | $0.30 | OSS |
| 17 |  MiniMax M2.5 | $0.12 | $0.99 | OSS |
| 18 |  Hunyuan A13B Instruct | $0.14 | $0.57 | OSS |
| 19 | Qwen3 235B A22B Thinking 2507 | $0.15 | $1.50 | OSS |
| 20 | Llama 4 Maverick | $0.15 | $0.60 | OSS |
| 21 |  LongCat Flash Chat | $0.20 | $0.80 | OSS |
| 22 | GPT-5 Mini | $0.25 | $2.00 | Closed |
| 23 | MiniMax M2 | $0.26 | $1.00 | OSS |
| 24 |  DeepSeek V3.2 | $0.26 | $0.38 | OSS |
| 25 |  Kimi K2.5 | $0.38 | $1.72 | OSS |
| 26 |  GLM 4.6 | $0.39 | $1.90 | OSS |
| 27 | GLM 4.7 | $0.39 | $1.75 | OSS |
| 28 | Qwen3.5 397B A17B | $0.39 | $2.34 | OSS |
| 29 | DeepSeek V3.2 Speciale | $0.40 | $1.20 | OSS |
| 30 | GPT-4.1 Mini | $0.40 | $1.60 | Closed |
Top 3 cheapest coding LLMs
gpt-oss-120b (free) is the cheapest model that clears our coding benchmark filter. At $0.00 per 1M input tokens, it lands well below frontier models while keeping credible SWE-bench and HumanEval scores.
gpt-oss-20b (free) follows close behind on price while holding up on coding-specific evals. Good fallback if gpt-oss-120b (free) is down or rate limited.
gpt-oss-120b follows close behind on price while holding up on coding-specific evals. Good fallback if gpt-oss-120b (free) is down or rate limited.
The price gap · cheapest vs most expensive
DeepSeek V3.2 Speciale is roughly Infinityx the price of gpt-oss-120b (free). The premium pays for long-context agentic coding, higher SWE-bench Verified scores, and vendor reliability. For one-shot code gen or refactors, the cheap options deliver 80 to 95 percent of the quality.