softperceptron.com / v0.1 — MVP live

The directory for
machine intelligence.

One place to discover AI tokens, deep learning models and autonomous agents — from free tiers to enterprise pricing. We link you straight to each provider. Payments, checkout and support happen on their site, not ours.

⚠ Disclaimer

SoftPerceptron is a directory only — we do not process payments, host the models, or vet provider checkouts. All external links, pricing and tools are operated by third parties and used entirely at your own risk. Verify the URL before paying.

Listings
25
Models
13
Agents
7

Three classes of intelligence.

/categories
/marketplace

Featured listings

Showing sample data — be the first to list something real.

ModelTop pick

Claude (Anthropic)

Claude Sonnet, Opus and Haiku — long-context reasoning, coding and agentic workflows.

Context
200k
Modalities
Text+Img
Models
Haiku/Sonnet/Opus
chatcodingagentsapi
Free → $3–$15 / Mtok
API + chat · Anthropic
View →
ModelFeatured

Gemini (Google)

Gemini 2.5 Pro / Flash — multimodal model with 1M-token context via AI Studio & Vertex.

Context
1M
Modalities
Text+Img+Audio+Video
Family
Pro / Flash / Nano
multimodallong-contextvertex
Free → $1.25–$10 / Mtok
API + app · Google DeepMind
View →
Model

GPT (OpenAI)

GPT-5, o-series reasoning and GPT-4o multimodal — the ChatGPT API.

Context
400k
Modalities
Text+Img+Audio
Tools
Realtime · Vision
chatreasoningrealtime
Free → $1.25–$10 / Mtok
API + ChatGPT · OpenAI
View →
Model

Grok (xAI)

Grok 4 — frontier reasoning with native web & X search.

Context
256k
Search
Native
Modalities
Text+Img
reasoningsearchapi
$3 → $15 / Mtok
API + Grok app · xAI
View →
ModelOpen

Llama (Meta)

Open-weight Llama 3.x / 4 family — download, fine-tune, self-host.

Sizes
1B → 405B
License
Llama Community
Format
safetensors
open-weightsfine-tuneself-host
Free (open weights)
weights + license · Meta AI
View →
Model

Mistral

Mistral Large, Medium, Codestral and open-weight Mixtral models.

Context
128k
Open
Yes (some)
Region
EU-hosted
europeanopencoding
Free → $2–$6 / Mtok
API + open weights · Mistral AI
View →
ModelCheap

DeepSeek

DeepSeek V3 and R1 — strong reasoning at a fraction of frontier prices.

Context
128k
License
MIT (R1)
Strength
Reasoning
reasoningopen-weightslow-cost
$0.14 → $2.19 / Mtok
API + open weights · DeepSeek
View →
Model

Cohere Command

Command R+ and Embed — enterprise RAG, retrieval and rerank.

Context
128k
Focus
RAG · Embed
Deploy
Cloud/Private
enterpriseragembeddings
$0.15 → $10 / Mtok
API · Cohere
View →
Model

Qwen (Alibaba)

Qwen3 — open-weight multilingual models, strong code & math.

Sizes
0.5B → 235B
Languages
100+
License
Apache-2.0
open-weightsmultilingualcode
Free (open weights)
weights + API · Alibaba
View →
Model

Midjourney

State-of-the-art text-to-image generation via Discord & web app.

Output
Images
Plans
Basic → Mega
Access
Web + Discord
imagegenerativeart
$10 → $120 / mo
subscription · Midjourney
View →
Model

Stable Diffusion

Stability AI — open-weight image, video and 3D generation models.

Family
SD3 / SDXL
Modalities
Image/Video/3D
License
Community/Comm.
imageopen-weightsvideo
Free → $20+ / mo
weights + API · Stability AI
View →
Model

ElevenLabs Voice

Realistic text-to-speech, voice cloning and dubbing in 70+ languages.

Languages
70+
Voices
5000+
Latency
<400ms
audiottscloning
Free → $330 / mo
subscription + API · ElevenLabs
View →
ModelOpen

Hugging Face Hub

1M+ open models, datasets and Inference Endpoints for any task.

Models
1M+
Datasets
250k+
Hosting
Spaces
openhubinference
Free → pay-as-you-go
hub + inference · Hugging Face
View →
TokenMulti-model

OpenRouter Credits

One API key, 300+ models — pay-as-you-go credits across providers.

Models
300+
Markup
0–5%
Auth
Single key
aggregatorcreditsbyok
From $5 top-up
credits · OpenRouter
View →
Token

Together AI Credits

Serverless & dedicated inference for 200+ open-source models.

Models
200+
GPUs
H100/H200
Trial
$1 free
inferenceopen-sourcecredits
$0.10 → $5 / Mtok
credits · Together AI
View →
TokenFastest

Groq Credits

Ultra-low-latency LPU inference for Llama, Mixtral and more.

Speed
500+ tok/s
Hardware
LPU
Free tier
Yes
low-latencylpucredits
Free → $0.05+ / Mtok
credits · Groq
View →
Token

Replicate Credits

Run thousands of open-source models with a simple HTTP API.

Models
10k+
Billing
Per-second
Deploy
Custom models
inferenceopendeploy
Pay-per-second
credits · Replicate
View →
Token

fal.ai Credits

Lightning-fast image, video and audio generative APIs.

Focus
Generative media
Latency
Sub-second
Free tier
Yes
imagevideocredits
Pay-per-call
credits · fal.ai
View →
AgentPopular

Perplexity

Answer engine — multi-step web research with citations & follow-ups.

Citations
Always
Modes
Pro · Deep Research
API
Sonar
researchsearchagent
Free → $20 / mo
subscription + API · Perplexity AI
View →
Agent

ChatGPT Agent

Autonomous browser & tool-using agent inside ChatGPT.

Tools
Browser · Code · Files
Plans
Plus · Pro · Team
Multimodal
Yes
agentbrowsertools
$20 → $200 / mo
Plus / Pro · OpenAI
View →
Agent

Manus

General-purpose autonomous agent that plans, browses and ships work.

Modes
Autonomous
Tools
Browser · Code
Output
Files · Sites
autonomousbrowseragent
Free → $39 / mo
subscription · Manus
View →
Agent

Devin

Cognition's autonomous software engineer — plans, codes and ships PRs.

Focus
Software eng.
Output
PRs · Repos
Integrations
GitHub · Slack
codingsweagent
From $20 / mo
ACU credits · Cognition Labs
View →
Agent

Cursor

AI-first code editor with multi-file agent that writes and refactors.

Plans
Free · Pro · Ultra
Models
Claude · GPT · Gemini
Mode
Agent · Compose
codingeditoragent
Free → $20–$200 / mo
subscription · Cursor
View →
Agent

Lindy

Build no-code AI employees for email, meetings, CRM and ops.

Triggers
200+
Tools
Gmail · Slack · CRM
No-code
Yes
no-codeworkflowagent
Free → $49.99+ / mo
subscription · Lindy
View →
Agent

ElevenLabs Conversational

Build production-grade voice agents with sub-second latency.

Latency
<400ms
Languages
32
SDKs
Web · iOS · Android
voiceagenttelephony
Free → usage-based
per minute · ElevenLabs
View →
/sell

List your model.
Send buyers to you.

Add your token, model or agent to the directory and link to your own pricing page. You handle checkout, delivery and support — we just send qualified traffic.

  • 01
    List in minutes
    Name, tagline, price label and a link to your page.
  • 02
    Keep 100%
    We never touch the transaction — you keep every cent.
  • 03
    Free for now
    Listing is free. Future ad placements support the directory.