AI Model Intelligence vs. Price: Find the Best Trade-off

August 10, 20262 min read

Choosing an AI model is rarely just about finding the highest score. A model can be excellent and still be a poor choice for everyday work if it costs too much. A cheaper model can be a great deal, but only if it is capable enough for the task.

This chart compares 100+ models using two dimensions:

  • AI Intelligence Index on the vertical axis: higher means stronger overall benchmark performance.
  • Blended price on the horizontal axis: an approximate cost per 1 million tokens, shown on a logarithmic scale.

Find the best intelligence for the price

Search a model, or click any point to pin its exact score and blended price.

104 models

No model pinned

Search or select a point to see its score and price.

$0.05$0.1$0.25$0.5$1$2$5$10$253035404550556065Cheap and SmartExpensive and SmartCheap and WeakExpensive and WeakBlended price · $/1M tokens (log)Intelligence IndexClaude Opus 5 (max)Claude Opus 5 (xhigh)Claude Fable 5 (with fallback)Claude Opus 5 (high)GPT-5.6 Sol (max)Kimi K3 (max)GPT-5.6 Sol (xhigh)Claude Opus 5 (medium)Qwen3.8 MaxGPT-5.6 Sol (high)Muse Spark 1.2 (xhigh)Grok 4.5 (high)GPT-5.6 Sol (medium)Claude Sonnet 5 (max)GPT-5.6 Terra (max)Claude Opus 4.7 (max)Muse Spark 1.1 (xhigh)GPT-5.4 (xhigh)Claude Opus 5 (low)GPT-5.6 Luna (max)Gemini 3.5 FlashDeepSeek V4 Flash 0731 (max)Gemini 3.6 FlashGPT-5.6 Sol (low)GPT-5.6 Terra (high)Gemini 3.1 Pro PreviewGPT-5.6 Terra (medium)Gemini 3.5 Flash (medium)GPT-5.6 Luna (high)MiniMax-M3DeepSeek V4 Pro (max)Kimi K2.6Claude Opus 4.6 (max)GPT-5.2 (xhigh)Kimi K2.7 CodeClaude Sonnet 5 (Non-reasoning)InklingDeepSeek V4 Flash (max)GPT-5.6 Sol (Non-reasoning)Nex-N2-ProGPT-5.2 Codex (xhigh)Inkling SmallGPT-5.4 mini (xhigh)Grok Build 0.1 0616Gemini 3 Pro Preview (high)Qwen3.6 PlusGPT-5.4 (low)GPT-5.6 Terra (low)JT-4.1 Flash 236B A21BQwen3.7 PlusDeepSeek V4 Flash (high)GPT-5.2 (medium)Claude Opus 4.6 (high)Gemini 3 FlashNemotron 3 UltraMiMo-V2.5GPT-5.6 Luna (medium)Grok 4.3 (high)GPT-4o Realtime (Dec)Qwen3.6 27BGemini 3.5 Flash-LiteGrok 4.3 (medium)Claude Sonnet 4.6 (Non-reasoning)GLM-5.1Gemini 3.5 Flash (minimal)GPT-5.5 (Non-reasoning)GPT-5.1 Codex (high)Claude Opus 4.5GLM 5V TurboClaude Sonnet 4.6 (Non-reasoning, Low Effort)GLM-5.2Claude Sonnet 5 (xhigh)Qwen3.5 27BGPT-5 (medium)Hy3-previewGPT-5.5 Instant (May 2026)Grok 4MiMo-V2-Flash (Feb 2026)GPT-5.4 Pro (xhigh)KAT-Coder-Pro V2Gemini 3 Pro Preview (low)o3-proGLM-5Qwen3.5 122B A10BQwen3.5 397B A17BDeepSeek V3.2Qwen3.6 35B A3BMiniMax-M2.1DeepSeek V4 ProMiMo-V2-FlashGPT-5 (low)GPT-5 mini (medium)Qwen3.5 Omni PlusGPT-5.1 Codex mini (high)o3Step 3.7 FlashDeepSeek V3.1 TerminusMistral Medium 3.5Kimi K2.5Claude 4.5 SonnetGLM-4.6JT-35B-FlashClaude 4.1 OpusMiMo-V2.5-Pro

Click a point or a search result to pin a model. The chart uses a logarithmic price axis so low-cost models remain visible.

How to read the chart

The dotted lines split the chart at an intelligence score of 50 and a blended price of $1 per million tokens:

  • Cheap and Smart: the most attractive area for high-volume work.
  • Expensive and Smart: strong models that may be worth the extra cost for difficult tasks.
  • Cheap and Weak: inexpensive models that can still be useful for simple work.
  • Expensive and Weak: models where the price is harder to justify from this comparison alone.

The horizontal axis is logarithmic. This keeps a model costing a few cents visible next to models costing several dollars or more. It also means that equal visual distances represent ratios rather than equal dollar differences.

A note about the data

This is a snapshot rather than a permanent ranking. Model versions, benchmark scores, provider pricing, and routing options change frequently. Prices are blended estimates, so the real cost for a particular workload will depend on the input/output mix and the provider you use.

The underlying comparison is based on the Artificial Analysis model index. Use it as a starting point for exploration, then check the latest model card and pricing before making a production decision.