Dashboard

The state of language models

Rankings, releases and benchmark movement · updated October 26, 2026
Models tracked
597
86 organizations
Open weights
389
65% of catalog
Releases · 90 days
49
tracked in the last quarter
Open–closed gap
379.8 Elo
MiMo-V2.6-Pro leads open
Quality vs. price
Frontier Elo against output price, log scale
Pareto frontier
050010001500200025003000$0.1$1$10$100Pareto frontier — best Frontier Elo at each priceClaude Opus 5.5Gemini 4 ArgonGPT-6.1 TerraKimi K3.1 ProMiMo-V2.6-ProMuse Wonder
Latest releases
Model explorer →
Oct 2026GPT-6.1 TerraOpenAI2997.6CLOSEDOct 2026Kimi K3.1 ProMoonshot AI3169.9CLOSEDOct 2026Nemotron 3.5 Bolt E4BNVIDIA1569.2OPENOct 2026iFlytek Spark X2.6 (Xinghuo X2.6)iFlytek1000.0CLOSEDOct 2026Spark-X2.6-4BiFlytek1760.2OPENOct 2026MiniMax M3.1 FlashMiniMax2745.6OPENOct 2026DeepSeek-V4.1-Flash-1017DeepSeek2747.3OPENOct 2026Muse Spark OpenMeta2595.7OPEN
Top ranked
All rankings →
1Kimi K3.1 Pro3169.9
2Gemini 4 Argon3126.1
3Muse Wonder Full3100.5
4Claude Opus 5.53072.4
5Muse Wonder3017.3
6GPT-6.1 Terra2997.6
7Claude Fable 5.12917.1
8GPT-6 Astra2882.1
Biggest movers
Elo gain over previous family release
Sarvam-105Bvs Sarvam-1 (2B)+1620.7Mistral Large 4vs Mistral Large 3+1227.7SmolLM3 3B (Think)vs SmolLM2 1.7B+1003.6DeepSeek-V4-Flash-0731vs DeepSeek-V4-Flash (Non-Think)+996.5Hy3vs Hunyuan-A13B+909.9
Quick compare