Serverless inference with pay-per-use token pricing ($0.0015 to $4.50 per 1M tokens for chat/vision models). Provisioned throughput with reserved PTUs for committed capacity. Dedicated GPU endpoints at $5.49-$8.99 per hour. Fine-tuning from $0.34-$40 per 1M tokens. GPU clusters at $1.99-$9.99 per GP
Free plan: $0 for up to 25 people with wikis, pages, records, notes, sharing, connected agents, CLI, sites, GitHub, 90 days of history, and limited chat. Advanced: $18 per person per month, includes Slack and Telegram, full history, and expanded chat allowance. Business: $27 per person per month, in
Key features
GPU clusters at scale with autoscaling
Voice agent infrastructure with sub-second latency
Comprehensive model library with 25+ regions
Browser-based wiki with history, diffs, notes and structured data tracking
Command-line interface for secure agent integration
Memory harness for Claude Code, Codex, Grok Build and Pi
Team chat integration with Slack, Microsoft Teams and Telegram
Conversational UI for company-wide knowledge access
Full audit trail with revision history and change attribution
Organization-level access control and wiki sharing
Single sign-on with Okta, Google Workspace and Microsoft Entra ID on Business plan
What makes it different
2x faster inference powered by cutting-edge research optimization
60% lower cost with workload-specific optimization
90% faster pre-training with Together Kernel Collection
Research-optimized infrastructure with proprietary kernels
Single unified API across multiple model providers
Agents automatically read organizational context before performing tasks and write back decisions
One unified knowledge source accessible across five different interfaces
Works with multiple AI agents and platforms without rebuilding context
Team members query knowledge in existing communication channels
Complete version control and attribution on all organizational changes