Model Intelligence
Leaderboards
Rank models by key performance metrics. Each leaderboard shows the full ranked list with provider, model type, and the primary score. Rankings are deterministic and derived from the Ethen normalized model index.
Intelligence
Composite benchmark score on the Ethen Intelligence Index. Higher scores indicate stronger overall model capability across reasoning, coding, and agentic tasks.
Speed
Measured output tokens per second. Higher values mean faster generation throughput.
Latency
Time to first answer token (TTFT). Lower values mean quicker initial response.
Input Price
Price per million input tokens. Lower values are more cost-effective for prompt-heavy workloads.
Output Price
Price per million output tokens. Lower values are more cost-effective for generation-heavy workloads.
Context Window
Maximum context window size. Larger windows enable processing of longer documents and conversations.
Also see: Benchmarks
Benchmark pages include a full explanation of each evaluation dimension alongside the ranked model table.
All Benchmarks →