LMSpeed

LMSpeed (lmspeed.net) helps you query model scores and test AI APIs before you choose. It aggregates per-token pricing from 100+ LLM API providers, so every model page shows a side-by-side table of input and output prices per million tokens and the cheapest provider for that model, next to benchmark signals such as Artificial Analysis overall, coding and math scores, context window, measured speed and first-token latency. The speed test runs five standardized prompts against any OpenAI-compatible endpoint (official providers, proxies, relays or self-hosted) and records first-token latency, output throughput and total duration. The all-in-one relay audit runs 13 security probes in one pass to check whether a relay swaps the model, injects hidden prompts, rewrites instructions, truncates context, substitutes tool calls, breaks streaming or leaks error details, then produces a redacted, shareable report; API keys are not written to public reports. LMSpeed also keeps a free LLM API directory (free tiers, free API keys and redeem codes for Claude, GPT, Gemini, DeepSeek, Llama, Qwen and more), a provider directory with health checks, and leaderboards for throughput, latency, provider health, pricing and the best models for agents, coding and reasoning.