Best intelligence per dollar

Quality index divided by reuse-adjusted blended cost at 50% reuse. Rewards models that are both capable and cheap to run once caching is working.

#ModelProviderQuality / $List blendedCache disc.
1OpenAI gpt-oss-120bFireworks372.9$0.2690%
2GPT-5 MiniOpenAI136.0$0.6990%
3Grok 4.3xAI78.7$1.5684%
4Kimi K2.6Fireworks60.0$1.7183%
5DeepSeek-V4-ProFireworks55.8$2.1792%
6GLM 5.1Fireworks49.9$2.1581%
7Claude Haiku 4.5Anthropic48.1$2.0090%
8Gemini 3.5 FlashGoogle30.0$3.3890%
9Gemini 3.1 Pro PreviewGoogle24.6$4.5090%
10Claude Sonnet 4.6Anthropic18.0$6.0090%
11Claude Opus 4.8Anthropic11.7$10.0090%
12Claude Opus 4.7Anthropic11.4$10.0090%
13GPT-5.5OpenAI10.0$11.2590%
14Claude Fable 5Anthropic5.9$20.0090%
15GPT-5.5 ProOpenAI1.4$67.500%
Method

intelligence ÷ reuseAdjustedBlended(model, 0.5, 0.25).

Rank models for your workload.

A diagnostic measures your real reuse and re-ranks the catalog for the way you actually call models.