Have an AI agent? Let it explore Baltor for you. Any assistant works, personal or coding.

Connect your agent
What your agent reads

The request it copies

Explore baltor.ai for me and report back. Start with https://baltor.ai/llms.txt, then read https://baltor.ai/.well-known/agent-card.json and https://baltor.ai/.well-known/mcp.json. Tell me which plan (Feeds at $4.99/mo, Pro at $29/mo, or Max at $49/mo) fits my work, what you would connect, and the setup steps at https://baltor.ai/connect. Stay read-only: read public pages and agent files, and check with me before any sign-in, purchase or change. If a step needs a key, name it (BALTOR_SERVICE_TOKEN) and I will add it myself.

Discovery files on baltor.ai

Read-only discovery is open. Keep BALTOR_SERVICE_TOKEN in your own environment.

Baltor
Sign in Get started

Can I run it?

Choose your hardware and see which open models and quantizations fit, at which context length.

What runs the model
Use your own numbers

What you enter stays on your computer.

On GeForce RTX 5090, 32 GB, 736 models fit at 8,192 tokens or more.

Read October 25, 2026
Models that fit on the device come first, then models that also use computer memory, with larger models first in each group. Speeds are estimates. Real results depend on the runtime, its settings and other work on the machine.
ModelQuantizationWeightsLongest context that fitsMemory thenWhereSpeed, estimate
Llama-4-Scout-17B-16E-InstructIQ2_XXS28.1 GiB8,19231.5 GiBOn the device135 to 217 tokens a second
Ling-flash-2.0IQ2_XXS25.3 GiB32,76829.0 GiBOn the device238 to 380 tokens a second
Solar-Open-100BIQ2_XXS25.3 GiB16,38430.1 GiBOn the device140 to 224 tokens a second
Qwen3-Next-80B-A3B-InstructQ2_K27.2 GiB16,38430.5 GiBOn the device249 to 399 tokens a second
Qwen3-Next-80B-A3B-ThinkingQ2_K26.2 GiB32,76831.0 GiBOn the device174 to 279 tokens a second

453 more need KV cache numbers from a source before the check can size them.

The estimateFormula
memory = weights + KV cache + overhead
weights = size of the quantized files,
or parameters × bits per weight ÷ 8
KV cache = 2 × layers × KV heads × head size
× context × bytes per value
overhead = 512 MiB + 5% of weights
speed = 50% to 80% of bandwidth
÷ (active weights + KV cache)

How the estimate works

For each model, the check picks the largest quantization that fits, then the longest context that fits with it. A model too large for the graphics card can still run with some layers in computer memory, more slowly.

Memory kept for the system

  • A graphics cardMay use all of its memory.
  • Unified memoryKeeps about a quarter to a third for the system.
  • Computer memory aloneKeeps a fifth for the system.

Enter your own memory to change these shares.

Where the numbers come from

Memory and bandwidth for each preset come from the vendor pages below, read September 24, 2026. Model sizes and architecture numbers come from each model's page. Add a bandwidth figure to any preset to see its speed.