The price spread between LLM APIs is now 100x. Groq Llama 8B costs $0.05/M input. GPT-5.4 Pro costs $30/M. Same prompt, wildly different bill.
I compiled pricing for every major model into one reference table.
Frontier Models (Best Quality)
| Model | Input/M | Output/M | Cache Hit/M | SWE-bench |
|---|---|---|---|---|
| DeepSeek V4 | $0.30 | $0.50 | $0.03 | 81% |
| GPT-5.4 | $2.50 | $15.00 | $0.25 | 80% |
| Claude Opus 4.6 | $5.00 | $25.00 | $0.50 | 80.8% |
| Claude Sonnet 4.6 | $3.00 | $15.00 | $0.30 | 79% |
| Gemini 3.1 Pro | $2.00 | $12.00 | $0.20 | 78% |
DeepSeek V4 is the outlier. Highest SWE-bench score at the lowest price. The catch: occasional outages and China data routing.
Mid-Tier (Best Value)
| Model | Input/M | Output/M |
|---|---|---|
| GPT-5.4 Mini | $0.75 | $4.50 |
| Claude Haiku 4.5 | $1.00 | $5.00 |
| Gemini 2.5 Flash | $0.30 | $2.50 |
| Mistral Large 3 | $2.00 | $6.00 |
Mistral Large 3 has the cheapest flagship output at $6/M — 60% less than GPT/Claude ($15/M).
Budget (Cheapest)
| Model | Input/M | Output/M |
|---|---|---|
| Groq Llama 8B | $0.05 | $0.08 |
| Gemini Flash-Lite | $0.10 | $0.40 |
| GPT-5.4 Nano | $0.20 | $1.25 |
| Mistral Small 3.1 | $0.20 | $0.60 |
What 10K Chatbot Replies/Day Actually Costs
| Model | Monthly Cost |
|---|---|
| Gemini Flash-Lite | $60 |
| DeepSeek V4 | $90 |
| GPT-5.4 Mini | $430 |
| Claude Sonnet 4.6 | $1,350 |
The full comparison covers 16+ models with cost-per-task breakdowns, hidden costs (long-context surcharges, data residency premiums), and a provider comparison (direct API vs gateway).
👉 Complete LLM pricing comparison table
Pricing from official provider pages. Cross-verified April 2026.
