Dashboard

The state of language models

Rankings, releases and benchmark movement · updated July 2026
Models tracked
537
84 organizations
Open weights
365
68% of catalog
Releases · 90 days
47
tracked in the last quarter
Open–closed gap
47.9 Elo
Kimi K3 leads open
Quality vs. price
Frontier Elo against output price, log scale
050010001500200025003000$0.1$1$10$100Claude Fable 5Claude Opus 4.8GPT-5.4 ProGPT-5.6Kimi K3
Latest releases
Model explorer →
Jul 2026Kimi K3Moonshot AIOPEN3025.3Jul 2026InklingThinking Machines LabOPEN2254.0Jul 2026Soofi SSOOFI ConsortiumOPEN764.0Jul 2026GPT-5.6OpenAICLOSED3073.2Jul 2026Muse Spark 1.1 (xhigh)MetaCLOSED2605.1Jul 2026Grok 4.5SpaceXAICLOSED2481.9Jul 2026Hy3TencentOPEN2485.4Jun 2026Claude Sonnet 5AnthropicCLOSED2490.3
Top ranked
All rankings →
1GPT-5.63073.2
2Kimi K33025.3
3Claude Fable 52936.7
4Claude Opus 4.82794.3
5GPT-5.4 Pro2786.0
6GPT-5.5 (xHigh)2725.9
7Claude Opus 4.72702.0
8GPT-5.52680.9
Biggest movers
Elo gain over previous family release
Sarvam-105Bvs Sarvam-1 (2B)+1547.2SmolLM3 3B (Think)vs SmolLM2 1.7B+975.4Hy3vs Hunyuan-A13B+968.3SmolLM3 3B (No Thinking)vs SmolLM2 1.7B+797.4Phi-4-reasoningvs Phi-4-mini (3.8B)+762.3
Quick compare