AI model comparison
Picking between two models usually comes down to three things: what a request really costs at your volume, how much context you get, and whether the model can call tools or read an image. Each page below lines those three up.
- GPT-5 vs Claude Sonnet 4.5
- GPT-5 vs Gemini 2.5 Pro
- GPT-5 Mini vs Gemini 2.5 Flash
- GPT-5 vs DeepSeek V3
- Claude Sonnet 4.5 vs Gemini 2.5 Pro
- DeepSeek V3 vs Qwen3 Coder 480B A35B
- GPT-5 Mini vs DeepSeek V3
- Grok 4.20 vs GPT-5
- Claude Haiku 4.5 vs GPT-5 Mini
- Gemini 2.5 Flash vs DeepSeek V3
- Qwen3 Max vs DeepSeek V3
Mistral Medium 3 vs GPT-5 Mini
Cheapest models, for a quick sanity check
| Model | ≈ 1,000 words |
|---|---|
| Mistral Nemo | 18 |
| Ling 3.0 Flash | 32 |
| Llama 3 8B Lunaris | 34 |
| MythoMax 13B | 45 |
| Nex-N2-Mini | 47 |
| Granite 4.0 Micro | 49 |
| Llama 3.1 8B Instruct | 49 |
| Mistral Small 3 | 49 |