Back to benchmarks

LLM Tier List

Rank this week's most-used paid models on OpenRouter alongside the Artificial Analysis leaders. Drag models into tiers, add any release from the searchable list, then export an image to share.

Model data updated Sep 25, 2026

Unranked 29Drag into a tier, or tap a model then a tier
DeepSeek V4.1 Flash
Opus 5.5
Fable 5.1
GLM 5.3 Flash
GPT-6 Sol
Hy4 preview
GPT-6 Astra
Qwen3.8 Max 0902
GPT-6 Luna
GLM 5.3
Grok 4.7
MiMo-V2.5
MiMo-V2.6-Pro
GPT-5.6 Terra
Gemini 3.8 Flash
Kimi K3
Qwen3.8 2.4T A95B
Muse Spark 1.3 Contributor
GPT-5.5
DeepSeek V4 Pro 0813
Qwen3.8 27B
Sonnet 5
Muse Spark 1.3
MiMo-V2.6-Flash
Solar Pro 4
Gemini 3.5 Flash Lite
MiniMax M3
Qwen3.8 Flash
Gemini 3.1 Pro Preview

How the pool is picked

Default merges every list below into one pool. Adding or removing a model yourself turns it into your own Custom pool. Most used ranks paid models by tokens on OpenRouter over the last 7 days; free variants don't count. Intelligence, Coding and Agentic use the Artificial Analysis indices published on OpenRouter's rankings.

One model per line

When a lab ships a new version of the same model line, only the newest makes the pool: Opus 5.5 replaces Opus 5, Grok 4.7 replaces 4.6. Distinct lines like GPT-6 Sol and GPT-6 Luna both stay. The search list syncs with OpenRouter daily, so you can rank a model the day it drops.