Skip to content
ModelCap

Model decision surface

Compare AI models

Start with Gemma 4 26B A4B and Gemma 4 31B, or choose any two current ranked language models. Compare capability evidence, price, context, provider availability, and weight access without pretending one field decides every use case.

Current public data

Gemma 4 26B A4B vs Gemma 4 31B

Live dataset updated 11/7/2026, 12:54:46 AM UTC

Open 1200×630 evidence receipt
Factual comparison of Gemma 4 26B A4B and Gemma 4 31B
Field
ModelCap position#56#37
Index score66.871.9
EvidenceMeasured2 public benchmark observations across 2 boardsMeasured2 public benchmark observations across 2 boards
Input / 1M$0.068$0.20
Output / 1M$0.225$0.40
Pricing statusfreshfresh
Context262K262K
Providers1515
Weight accessOpen weightsOpen weights

Decision facts

  • Gemma 4 26B A4B is #56; Gemma 4 31B is #37 on the same current language board.
  • Index scores are 66.8 for Gemma 4 26B A4B and 71.9 for Gemma 4 31B. Their published uncertainty intervals overlap, so the rank difference alone does not establish a reliable capability advantage for your workload.
  • Both positions use Measured evidence.
  • Listed output price per 1M tokens is $0.22 for Gemma 4 26B A4B and $0.40 for Gemma 4 31B. For 1,000 requests with 2,000 input and 500 output tokens each (2M input + 0.5M output), the listed-rate estimate is $0.25 for Gemma 4 26B A4B versus $0.60 for Gemma 4 31B. Gemma 4 26B A4B costs 58.8% less in this scenario. This excludes caching, batch discounts, prompt-length tiers, tool charges and retries; verify the selected endpoint before budgeting.
  • Published context is 262,144 tokens for Gemma 4 26B A4B and 262,144 for Gemma 4 31B.

These are separate published fields, not a synthetic winner. ModelCap does not collapse price, access, context, and capability evidence into a hidden recommendation score.

Each comparison page is a permanent, shareable URL with the same live figures as this tool: ModelCap Index position, API pricing, context window, provider count, weight access and every shared benchmark board.