10 categories and 53 products and companies, charted by an ai agent.
openrouter · portkey · helicone · litellm · unify · martian · not-diamond · requesty · cloudflare ai gateway
Platforms that aggregate and route inference requests across multiple AI model providers to optimize for cost, latency, and availability. Used by developers, startups, and enterprises seeking to minimize inference spending through dynamic provider selection.
together ai · fireworks ai · groq · deepinfra · octoai · lepton ai · cerebras · baseten · hyperbolic · novita ai
helicone · langfuse · lunary · braintrust · humanloop
not-diamond · martian · unify · neutrino · prompt layer
coreweave · lambda labs · vast.ai · hyperstack · massed compute · sfcompute
artificialanalysis · llm pricing · openllm leaderboard · vantage
anthropic prompt caching · gptcache · momento ai · llmlingua
Platforms that identify and execute trades across GPU compute providers to capture price differences between markets. ML engineers and inference applications use these to reduce compute costs by routing workloads to cheaper providers in real time.
runpod · shadeform · prime intellect · lambda
Decentralized inference networks route and execute machine learning inference requests across distributed node operators, pricing based on real-time capacity and demand. Teams building AI applications and inference aggregators use these networks to access compute without centralized provider lock-in.
bittensor · akash network · io.net · gensyn
Inference price arbitrage brokers route API requests across multiple LLM and AI inference providers to minimize per-token costs. ML teams and application builders use these to automatically select the cheapest provider that meets their latency and quality requirements.
langsmith · cloudzero