Home/Models/Model finder
Source-tracked model selection

Find AI models that fit your workload and budget.

Filter by practical requirements, estimate monthly token cost, and open a direct comparison before you test a model in your own application.

Explore the price mapOpen the full calculator
Your requirements

Describe the decision

Monthly workload
Options to evaluate

Lower-cost matches for general text or chat

Showing 5 of 43 models.

How to read this shortlist: fit signals reflect your selected requirements and tracked data. They are not quality scores or success probabilities.
01Stable

Groq · GPT-OSS

GPT-OSS 20B on Groq

$2.25 / month

Input / 1M
$0.075
Output / 1M
$0.3
Context
131,072
Cost Lower in setTask Check fitData Tracked

Matches the selected availability and pricing requirements. Test quality and latency with your own workload.

Model detailsPricing source
03Stable

Google · Gemini 2.5

Gemini 2.5 Flash-Lite

$3.00 / month

Input / 1M
$0.1
Output / 1M
$0.4
Context
1,048,576
Cost CompareTask Check fitData Tracked

Matches the selected availability and pricing requirements. Test quality and latency with your own workload.

Model detailsPricing source
04Stable

Cohere · Command R

Command R 08-2024

$4.50 / month

Input / 1M
$0.15
Output / 1M
$0.6
Context
128,000
Cost CompareTask Check fitData Tracked

Matches the selected availability and pricing requirements. Test quality and latency with your own workload.

Model detailsPricing source
05Stable

OpenAI · GPT-4o

GPT-4o mini

$4.50 / month

Input / 1M
$0.15
Output / 1M
$0.6
Context
128,000
Cost CompareTask Check fitData Tracked

Matches the selected availability and pricing requirements. Test quality and latency with your own workload.

Model detailsPricing source
Validate the shortlist

Compare the first two options with the same workload

Cost is calculated from tracked token rates. It does not assume equal quality, latency, retry rate, or output length.

Compare these models
How recommendations work

Requirements first, then a factual sort

StackLens filters models using tracked availability, capabilities, modalities, context windows, and pricing. It then sorts the matching set by your selected cost, context, or caching preference.

Read the StackLens methodologyReview every tracked model
Decision boundary

A shortlist is not a benchmark winner

These are options worth testing, not claims about model quality. Evaluate task accuracy, latency, retries, regional availability, safety behavior, and provider terms before production use.

Explore model relationshipsReview the monthly cost guide