llm.ing

GPT-5

OpenAIOpenRouter ↗

Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows

Benchmarks

BenchmarkCategoryScoreVariantProvenanceSourceObserved
AA Agentic Indexagentic26.5—mirroropenrouterAug 7, 2026
26.5—independentaaAug 7, 2026
AA Coding Indexcoding37.8—mirroropenrouterJul 23, 2026
37.8—independentaaJul 24, 2026
AA Intelligence Indexintelligence23.0—independentaaSep 9, 2026
23.0—mirroropenrouterOct 5, 2026
Design Arena — 3dpreference1028—mirroropenrouterNov 19, 2026
Design Arena — asciiartpreference1124—mirroropenrouterNov 17, 2026
Design Arena — codecategoriespreference1152—mirroropenrouterNov 19, 2026
Design Arena — datavizpreference1221—mirroropenrouterOct 17, 2026
1217—mirroropenrouterNov 19, 2026
Design Arena — gamedevpreference1168—mirroropenrouterNov 19, 2026
Design Arena — svgpreference1174—mirroropenrouterNov 19, 2026
Design Arena — uicomponentpreference1166—mirroropenrouterNov 19, 2026
Design Arena — websitepreference1173—mirroropenrouterNov 18, 2026
LMArena Textpreference1406high-effortcrowdlmarenaOct 15, 2026
LMArena Visionpreference1208high-effortcrowdlmarenaOct 9, 2026
Median output speedspeed113 tok/s—independentaaSep 22, 2026
Median time to first tokenspeed66.01s—independentaaSep 22, 2026

Includes index data from Artificial Analysis.

Providers

Live operational stats per hosting endpoint, via OpenRouter.

ProviderQuantContext$/1M in · outUptime 24hLatencyThroughput
Azure—400K$1.25 · $10100.00%——
Azure—400K$1.38 · $11100.00%——
OpenAI—400K$1.25 · $10100.00%——