Topic

Model leaderboards

Public ranked comparisons of language models on shared task sets, usually maintained by third parties and reported as a single ordering.

Current clusters

build1 publisher

Two open-weight launches clear frontier-tier numbers on vendor engineering claims

DeepSeek's V4.1 Flash claims a million-token context on an FP4 KV cache and Alibaba opened 2.4 trillion parameters. The seven reports collected around the two launches quote no latency, throughput or hardware.

Publishers:dev.to

Reality

Evidence28
Adoption12
Hype gap+40
Incentives72
Confidence35