TokenSave logo TokenSave

All comparisons · Scores checked 2026-11-18 · Prices 2026-10-14

AI model capability vs price

A more expensive model is not always a smarter one. This chart puts each model's general capability, measured by the Epoch Capabilities Index (ECI), against what its API costs. The highest score is Claude Fable 5.2 at 173.5.

Best value
Claude Opus 5.5
ECI 168.7 · $11.04 per 1M tokens

4.8 points below the top model for 50% of its price, and on the best-value frontier.

Scatter chart of AI model capability (Epoch Capabilities Index) against API price per million tokens, with the best-value frontier highlighted
Capability (ECI) vs blended API price per 1M tokens, 31 models. Frontier models are the best value at their price.
⏳ Waiting for a score

These models are new, and Epoch AI has not published a capability score for them yet. They join the charts automatically once it does. Their prices today:

  • GPT-6.2 Astra$8 / $40 · $16.00
  • GPT-6.2 Sol$2 / $10 · $4.00
  • Mistral Large 3$0.5 / $1.5 · $0.79

The best-value models

A model is on the best-value frontier when no other model is both cheaper and more capable. From cheapest to most capable: GPT-5 nano, GPT-6 Luna, Gemini 3.8 Flash, GPT-6.1 Terra, GPT-6.1 Sol, Claude Opus 5.5, GPT-6 Astra (legacy), Claude Fable 5.2. Anything below the line costs more than a frontier model with the same or higher score.

All scores and prices

ModelECI
(90% range)
Input / output
per 1M
Blended
per 1M
Claude Fable 5.2 ★ best value173.5
170–177
$8 / $40$22.08
GPT-6 Astra (legacy) ★ best value169.2
166–172
$10 / $50$20.00
Claude Opus 5.5 ★ best value168.7
166–172
$4 / $20$11.04
GPT-6.1 Sol ★ best value168.6
166–172
$2 / $10$4.00
GPT-6.1 Terra ★ best value167.7
165–171
$1 / $5$2.00
Claude Fable 5.1 (legacy)167.1
164–170
$10 / $50$27.60
Claude Sonnet 5.5166.6
164–170
$2 / $10$5.52
Gemini 4 Argon166.6
164–170
$2 / $10$4.00
GPT-6 Sol163.8
161–167
$2 / $10$4.00
GPT-5.5160.2
158–162
$5 / $30$11.25
Gemini 3.8 Flash ★ best value158.8
157–161
$0.75 / $3.75$1.50
Kimi K3158.6
157–161
$3 / $15$6.00
Muse Spark 1.3158.2
156–160
$1.25 / $4.25$2.00
Grok 4.7156.4
154–159
$2 / $6$3.00
GPT-6 Luna ★ best value156.4
154–158
$0.1 / $0.5$0.20
Gemini 3.1 Pro156.2
154–158
$2 / $12$4.50
Qwen 3.8 Max156.0
154–158
$2 / $6$3.00
DeepSeek V4 Pro156.0
154–158
$1.32 / $3.96$1.98
DeepSeek V4.1 Flash154.4
153–156
$0.3 / $1.2$0.52
Grok 4.20152.3
151–154
$1.25 / $2.5$1.56
Kimi K2.6151.6
150–153
$0.95 / $4$1.71
GPT-5 mini146.0
144–148
$0.25 / $2$0.69
Gemini 3.5 Flash-Lite145.7
144–147
$0.3 / $2.5$0.85
Gemini 3.1 Flash-Lite145.0
143–147
$0.25 / $1.5$0.56
Claude Haiku 4.5142.8
141–144
$1 / $5$2.06
Mistral Medium 3.5141.6
138–145
$1.5 / $7.5$3.15
GPT-5 nano ★ best value139.1
137–141
$0.05 / $0.4$0.14
GPT-4.1136.2
134–138
$2 / $8$3.50
GPT-4.1 mini134.3
132–136
$0.4 / $1.6$0.70
GPT-4o128.8
126–131
$2.5 / $10$4.38
GPT-4o mini123.8
120–127
$0.15 / $0.6$0.26

Blended price = (3 × input + output) ÷ 4 per million tokens, a typical chat mix, adjusted for each model's tokenizer (how many tokens it uses for the same English text). Standard list prices, no caching or batch discounts.

What the score means

ECI combines dozens of benchmarks (math, coding, science, reasoning and more) into one scale, so models tested on different benchmarks can still be compared. A gap of a few points is often within the margin of error; the 90% range in the table shows how sure the estimate is. It measures general capability, not your task: for a specific job, test a few frontier models on your own prompts.

To see what your own prompts cost on each model, paste them into the token counter, or compare two models side by side on the comparison pages.

What the chart means, model by model: Best value LLM in October 2026.

Capability scores: Epoch Capabilities Index by Epoch AI, used under CC BY 4.0. Prices: TokenSave, updated daily.