Best-value AI models
The frontier contains models for which no eligible alternative is at least as capable and no more expensive, with a strict advantage in one of those measures. It helps you compare trade-offs between benchmark quality and cost.
Which should you choose?
No single benchmark covers every use case. Try your own prompts and compare accuracy, response time, input/output support, and the actual bill.
AI models
| Model | Intelligence score | Benchmark cost (USD) | Input price / 1M tokens (USD) | Output price / 1M tokens (USD) | Output speed |
|---|---|---|---|---|---|
| GPT-5.6 Luna (Non-reasoning)OpenAIBest value frontier | 26.80 | $10.2 | 0.20 | 1.20 | 163 tok/s |
| GPT-5.6 Luna (low)OpenAIBest value frontier | 33.90 | $13.7 | 0.20 | 1.20 | 167 tok/s |
| GPT-5.6 Luna (medium)OpenAIBest value frontier | 38.90 | $21.1 | 0.20 | 1.20 | 179 tok/s |
| GPT-5.6 Luna (high)OpenAIBest value frontier | 47.00 | $55.0 | 0.20 | 1.20 | 177 tok/s |
| GPT-5.6 Luna (xhigh)OpenAIBest value frontier | 50.10 | $95.1 | 0.20 | 1.20 | 178 tok/s |
| Gemini 3.7 Flash (low)GoogleBest value frontier | 50.90 | $165 | 0.75 | 3.75 | - |
| GPT-5.6 Luna (max)OpenAIBest value frontier | 52.30 | $172 | 0.20 | 1.20 | 177 tok/s |
| Gemini 3.7 Flash (medium)GoogleBest value frontier | 53.40 | $278 | 0.75 | 3.75 | - |
| Gemini 3.7 Flash (high)GoogleBest value frontier | 56.00 | $485 | 0.75 | 3.75 | - |
| Muse Spark 1.2 (xhigh)MetaBest value frontier | 56.80 | $639 | 1.25 | 4.25 | - |
| GPT-5.6 Sol (high)OpenAIBest value frontier | 57.30 | $955 | 5.00 | 30.00 | 70 tok/s |
| Grok 4.6 (high)SpaceXAIBest value frontier | 60.90 | $1068 | 2.00 | 6.00 | 58 tok/s |
| Claude Opus 5 (Adaptive Reasoning, High Effort)AnthropicBest value frontier | 61.50 | $1974 | 5.00 | 25.00 | 58 tok/s |
| Claude Opus 5 (Adaptive Reasoning, Xhigh Effort)AnthropicBest value frontier | 62.50 | $2909 | 5.00 | 25.00 | 54 tok/s |
| Claude Opus 5 (Adaptive Reasoning, Max Effort)AnthropicBest value frontier | 63.10 | $3836 | 5.00 | 25.00 | 53 tok/s |
Guides
How we compare models
Intelligence uses the Artificial Analysis Intelligence Index. Benchmark cost is the total USD cost of running that test, including reasoning tokens; it is not a per-token API price. Rankings use the same dataset and calculations in every language. General benchmark scores do not measure performance in a particular language or task.