Analyzes how model pricing distributes across 300 AI models and which providers offer the best deals. Compare input vs output costs, price tiers, and find the highest-scoring model at every budget.
| Tier | Count | Avg Score |
|---|---|---|
| Free | 23 | 65 |
| Ultra-Budget | 109 | 66 |
| Budget | 79 | 69 |
| Mid-Range | 70 | 73 |
| Premium | 9 | 66 |
| Enterprise | 10 | 76 |
I/O Ratio = output price / input price. A ratio of 3.0x means output tokens cost 3x more than input tokens.
| Model | $/M In | $/M Out | Ratio |
|---|---|---|---|
| Qwen3 Next 80B A3B Instruct | $0.09 | $1.10 | 12.2x |
| Qwen3 VL 30B A3B Thinking | $0.13 | $1.56 | 12.0x |
| Qwen3 VL 8B Thinking | $0.12 | $1.36 | 11.7x |
| Palmyra X5 | $0.60 | $6.00 | 10.0x |
| Qwen3 VL 235B A22B Thinking | $0.26 | $2.60 | 10.0x |
| Qwen3 235B A22B Thinking 2507 | $0.15 | $1.50 | 10.0x |
| Nova 2 Lite | $0.30 | $2.50 | 8.3x |
| Gemini 2.5 Flash | $0.30 | $2.50 | 8.3x |
| Seed-2.0-Lite | $0.25 | $2.00 | 8.0x |
| GPT-5.3 Chat | $1.75 | $14.00 | 8.0x |
| Model | $/M In | $/M Out | Ratio |
|---|---|---|---|
| Molmo2 8B | $0.20 | $0.20 | 1.0x |
| Rnj 1 Instruct | $0.15 | $0.15 | 1.0x |
| Ministral 3 14B 2512 | $0.20 | $0.20 | 1.0x |
| Ministral 3 8B 2512 | $0.15 | $0.15 | 1.0x |
| Ministral 3 3B 2512 | $0.10 | $0.10 | 1.0x |
| Cogito v2.1 671B | $1.25 | $1.25 | 1.0x |
| Spotlight | $0.18 | $0.18 | 1.0x |
| Llama Guard 4 12B | $0.18 | $0.18 | 1.0x |
| Qwen2.5 VL 72B Instruct | $0.80 | $0.80 | 1.0x |
| R1 Distill Qwen 32B | $0.29 | $0.29 | 1.0x |
The highest-scoring model at or below each average cost threshold.
| Max $/1M | Model | Score | Actual Cost |
|---|---|---|---|
| $0.10 | Nemotron 3 Super (free) | 84 | Free |
| $0.50 | Grok 4.1 Fast | 87 | $0.35/1M |
| $1.00 | Grok 4.1 Fast | 87 | $0.35/1M |
| $2.00 | Gemini 3 Flash Preview | 89 | $1.75/1M |
| $5.00 | GPT-5.4 Mini | 93 | $2.63/1M |
| $10.00 | GPT-5.4 | 94 | $8.75/1M |
| $20.00 | GPT-5.4 | 94 | $8.75/1M |
| $50.00 | GPT-5.4 | 94 | $8.75/1M |
| Provider | Models | Avg Cost |
|---|---|---|
| Liquid AI | 5 | $0.02 |
| StepFun | 2 | $0.10 |
| Allen AI | 7 | $0.24 |
| NVIDIA | 11 | $0.33 |
| Baidu | 5 | $0.45 |
| Meta | 13 | $0.48 |
| Inception | 3 | $0.50 |
| ByteDance | 5 | $0.57 |
| arcee-ai | 7 | $0.57 |
| Alibaba | 51 | $0.62 |
| MiniMax | 8 | $0.67 |
| DeepSeek | 11 | $0.68 |
| Xiaomi | 3 | $1.13 |
| Mistral AI | 25 | $1.16 |
| Moonshot AI | 4 | $1.28 |
| 23 | $1.90 | |
| Amazon | 5 | $2.23 |
| aion-labs | 3 | $2.75 |
| Cursor | 2 | $3.00 |
| Cohere | 4 | $3.24 |
| xAI | 10 | $3.73 |
| Perplexity | 5 | $5.80 |
| Inflection | 2 | $6.25 |
| Anthropic | 13 | $14.55 |
| OpenAI | 59 | $18.54 |
We categorize models into tiers based on average cost per million tokens: Free (zero cost), Budget (under $1), Mid ($1-$10), Premium ($10-$50), and Enterprise (over $50). This tiering helps developers quickly find models that match their budget constraints.
The I/O ratio compares the cost of output tokens to input tokens. A ratio of 3x means output costs three times more than input. This matters because output-heavy workloads (like content generation) will cost more with high-ratio models, while input-heavy workloads (like analysis) are less affected.
Use the "Best Value at Each Price Point" table to find the highest-scoring model within your budget. Set your maximum cost threshold, and the table shows which model delivers the best quality at or below that price. You can also check the scatter chart to visually identify models with the best score-to-cost ratio.