Best AI Models by Price and Performance

The most SI Score per dollar, using a blended price of input and output tokens at a 3:1 mix.

updated Oct 8, 2026 · 15 models ranked

How this list is ranked: Ranked by value score = SI Score ÷ (blended price + $0.25), where blended price = 0.75 × input + 0.25 × output per 1M tokens (a typical 3:1 mix). The $0.25 floor keeps free tiers from exploding the ratio.

#	Model	Value (SI per $)	SI Score	Confidence	Price in	Context
1	Claude Haiku 5.5 Anthropic	144	64.8	100% confidence 100 percent, Full	$0.10	1M
2	GPT-6 Luna OpenAI	136.9	61.6	100% confidence 100 percent, Full	$0.10	1.1M
3	DeepSeek V4.1 Flash DeepSeek	129.2	66.2	69% confidence 69 percent, Medium	$0.15	1M
4	DeepSeek V4 Pro 0813 DeepSeek provisional	50.8	63.0	69% confidence 69 percent, Medium	$0.66	1M
5	Gemini 3.5 Flash Lite Google	50.2	55.2	100% confidence 100 percent, Full	$0.30	1M
6	Gemini 3.7 Flash Google	40	70.0	100% confidence 100 percent, Full	$0.75	1M
7	Gemini 3 Flash Preview Google provisional	39.6	54.4	85% confidence 85 percent, High	$0.50	1M
8	Gemini 3.8 Flash Google	39.3	68.8	100% confidence 100 percent, Full	$0.75	1M
9	Gemini 3.6 Flash Google	37.5	65.6	100% confidence 100 percent, Full	$0.75	1M
10	Grok 4.7 xAI	20.2	65.5	100% confidence 100 percent, Full	$2.00	500K
11	Gemini 3.5 Flash Google	18.8	68.3	100% confidence 100 percent, Full	$1.50	1M
12	GPT-6.1 Sol OpenAI	17.2	73.0	100% confidence 100 percent, Full	$2.00	1.1M
13	Claude Sonnet 5.5 Anthropic	16.3	69.4	93% confidence 93 percent, High	$2.00	1M
14	Gemini 3.1 Pro Preview Google provisional	14.3	68.1	100% confidence 100 percent, Full	$2.00	1M
15	Gemini 2.5 Pro Google	12.3	45.5	85% confidence 85 percent, High	$1.25	1M

Task pages rank on one transparent metric, not the blended SI Score — see methodology for how pillars and confidence are computed. Missing values mean the source has not reported them.