LLM API pricing comparison
Compare AI model cost by use case.
Check token price, context size, and simple cost estimates for chatbots, RAG, summarization, and coding agents. The site uses clear English and public pricing data.
Catalog last generated: Jul 13, 2026. Check data sources before production decisions.
Database paths
Start from the evidence layer
Model database
Browse normalized pricing, token limits, capability flags, and source links by provider route.
Providers
Browse model pricing by provider. Compare model counts, price ranges, and review route-level pricing per platform.
Local LLMs
Explore open-weight models with VRAM requirements, parameter counts, and quantization support for local inference.
Exact model pricing
Jump from exact model and route pricing searches into current catalog rows and detail pages.
Benchmark explorer
Compare source-backed benchmark rows within one benchmark instead of mixing unrelated scores.
Data sources
Check how the catalog is generated, what needs provider verification, and where corrections belong.
Use-case guides
Pick a model by the job you need to do
Chatbot
Fast replies, low cost per message, and easy scaling for support or product chat.
RAG
Large context and stable token cost for answers built from your own docs.
Summarization
Low-cost reading of long files, articles, or notes with short output.
Coding agent
Compare models for long prompts, repeated turns, and higher output volume.
Featured articles
Read the data before you compare
Cheapest LLM API chat models for a 500-token chatbot workload
Start with a 100k-message chatbot cost screen, then jump into model pages and compare links from current pricing data.
Cheapest LLM API models for coding workloads
Compare low-cost models for a coding workload with 2k input and 1.5k output tokens per task.
GPT-5.4 Mini vs Gemini 3.1 Flash Lite vs Claude Haiku 4.5
A data-led comparison of three compact models, focused on output cost, context, and capability tradeoffs.
LLM API cache and batch pricing guide
Learn when cached input, batch, priority, and flex pricing rows change the cost picture.
LLM API Provider Pricing Comparison
Compare OpenAI, Anthropic, Google, Mistral, and DeepSeek across cheapest chat, mid-range, and reasoning model pricing.
Model Routing Cascade
A four-tier Budget, Mid, Premium, and Reasoning model selection framework for routing queries to the lowest-cost suitable tier.
Agentic AI cost estimation guide
Multi-step agent costs across budget, cascade, and premium strategies. Context accumulation, tool calls, reasoning multipliers, and cache optimization.
Cheapest LLM API embedding models for vector search
Compare low-cost embedding models for a 500-token vector-index workload. See pricing, dimensions, and per-document costs.
Hidden costs of LLM APIs beyond per-token pricing New
Fine-tuning, rate limits, latency tradeoffs, evaluation, integration, and vendor risk — the LLM API costs that don't show up in per-token comparisons.
Output vs Input pricing: why generation costs 3–4× more New
Output tokens cost 3.6× more than input on average across 2,076 chat models. Provider breakdowns, reasoning vs non-reasoning, and real workload estimates.
Local LLMs vs cloud APIs: cost and use-case comparison New
Compare open-weight local models against cloud API pricing. Learn when local deployment wins, when cloud is better, and hardware requirements.
Reasoning LLM API pricing guide New
Which reasoning models are cheapest, which are most expensive ($168/1M output), and when to use reasoning vs non-reasoning models — with cached pricing data for 466 models.
LLM API cost cascade: how model routing saves up to 90% New
Budget models cost 50-200x less than premium for simple tasks. A data-led comparison of 745 budget vs 247 premium models with a routing framework to capture the savings.
Open-weight inference provider pricing comparison New
Compare pricing across 11 open-weight inference providers. Groq, Fireworks, DeepInfra, Together AI — cheapest models, near-1:1 output/input ratios.
Price structure
Input vs output price
Linear axes show absolute per-token price differences for chat models. Use the calculator for workload-specific estimates.
Capability metadata
Capability coverage
Based on LiteLLM metadata flags, not independent feature verification.
Benchmarks
Collected benchmark snapshots
Benchmark coverage is intentionally explicit and source-linked.
Need exact costs for your workload?
Use the calculator for custom token counts, request volume, and cache assumptions.