Directories/Models/Value frontier

The model value frontier

Compare published intelligence scores with listed API prices. The line shows the highest score at each price in this chart. Benchmark performance is one signal; test the models on your own work.

Full ranking board →What changed today

500 plotted records out of 599 with an intelligence score · maximum 500 records · Intelligence Index from Artificial Analysis · check source dates on model pages

Intelligence per dollar

Up and to the left is better: smarter and cheaper. The frontier shows the price and score trade-off among these records. Capabilities, reliability and your task results also matter.

value frontierOpenAIQwenAnthropicGoogleDeepSeekZ AiNVIDIAMoonshotai

Best-value picks

3 on the frontier

Model	Intelligence	Blended $/M	In / Out $/M
Claude Opus 5.5 (batch)Anthropic	57.6	$4	$2 / $10
Claude Sonnet 5.5 (batch)Anthropic	56	$2	$1 / $5
Claude Sonnet 5 (Free)Anthropic	53.4	$0	$0 / $0

Fastest models

tokens/sec

Mercury 2.5738 t/s Gemini 3.5 Flash Lite447 t/s Gemini 3.5 Flash Lite (batch)447 t/s Mercury 2442 t/s Nemotron 3.5 Super VL413 t/s Nemotron 3.5 Super VL (free)413 t/s Muse Spark 1.2380 t/s Gemini 2.5 Flash-Lite375 t/s

Median output speed from Artificial Analysis. Higher is faster to stream a response.

The context-window race

top 12

Grok 4.1 Fast2M Grok 4 Fast2M Grok 4 Fast2M Grok 4.202M GPT-6 Astra1.1M GPT-6 Astra (batch)1.1M GPT-6.1 Sol1.1M GPT-6.1 Sol (batch)1.1M GPT-6 Sol1.1M GPT-6 Sol (batch)1.1M GPT-5.6 Sol1.1M GPT-5.6 Sol (batch)1.1M