LL
Map
Models
Hardware
Providers
Benchmarks
Guides
News
Insights
Search…
⌘K
Menu
Models
Hardware
Providers
Benchmarks
Guides
News
Insights
Catalogue
Models
376 models, cross-linked to hardware verdicts and live provider pricing.
Text
287
Image
11
Audio
78
All weights
Open
Open (restricted)
Hosted only
All vendors
OpenAI (59)
Qwen (48)
Google (26)
Mistral AI (21)
NVIDIA (18)
Anthropic (16)
Z.AI (12)
DeepSeek (9)
Meta (8)
MiniMax (8)
Moonshot AI (7)
IBM (7)
OpenRouter (6)
Amazon (5)
xAI (5)
Meta (5)
Perplexity (5)
Microsoft (5)
Nous Research (4)
ByteDance Seed (4)
Useful Sensors (4)
TheDrummer (4)
AionLabs (4)
Sao10K (3)
Tencent (3)
Kwaipilot (3)
poolside (3)
AutoArk AI (3)
ESPnet (3)
inclusionAI (3)
Any size
< 8B
8–30B
30–100B
100B+
Any context
≥ 64K
≥ 128K
≥ 1M
Sort: newest first
Sort: largest first
Sort: name A–Z
1 model
Model & lab
The model
Released
Params
Context
Weights
Cheapest host
In / out per 1M
Speed
Compare
P
Perceptron Mk1
Perceptron
Perceptron Mk1 is a hosted-only system that reads text, images and video and writes text back, with a 32,768-token limit per request.
May 2026
—
33K
HOSTED ONLY
$0.15 / $1.50
per 1M
via OpenRouter
—