Rank Changes
Follow the biggest movements in the AI model landscape. These movers are calculated against the latest weekly snapshot rather than simulated historical jitter.
Baseline snapshot: Sep 14, 2026 to Sep 20, 2026.
Gainers
Claude Opus 5.5
Anthropic · coding
Grok 4.7
xAI · coding
MiMo-V2.6-Pro
Xiaomi · coding
Claude Opus 5
Anthropic · coding
Claude Sonnet 5
Anthropic · coding
gpt-oss-20b (batch)
OpenAI · coding
Qwen3.8 Max Prime
Alibaba · coding
Aion 3.5 Mini
aion-labs · coding
Aion 3.5
aion-labs · coding
Solar Mini 4
Upstage · coding
Losers
Ternary Bonsai 2 27B
prism-ml · coding
Pareto
unbiased · coding
DeepSeek Pro Latest
~deepseek · coding
DeepSeek Flash Latest
~deepseek · coding
Schematron V2 Turbo
inference-net · coding
Schematron V2 Small
inference-net · coding
GPT Astra Latest
~openai · coding
GPT Sol Latest
~openai · coding
GPT Terra Latest
~openai · coding
GPT Luna Latest
~openai · coding
Hot
New
Use This With
Rank changes are calculated by comparing current live rankings against the most recently archived weekly snapshot. A positive value means the model moved up (improved), while a negative value means it dropped.
Rankings update hourly based on live data from a multi-source pipeline. Weekly snapshots are archived at the end of each week and serve as the baseline for the next week of change calculations.
The Hot section highlights models that have gained significant attention or usage growth recently, even if their composite rank movement is modest. It helps surface emerging trends.
New entrants are models that appeared in our tracking system for the first time during the most recent snapshot period. They are typically newly launched endpoints from providers or freshly onboarded models.