Bindu Reddy Says Ox-Alpha Underperforms Older Kimi Model
Abacus.AI CEO Bindu Reddy shares benchmarks claiming Ox-Alpha lags behind older Kimi model.
Bindu Reddy posted that her team evaluated Ox-Alpha and found its scores worse than Kimi 2.6, a model two generations older. She called the surrounding attention pure marketing despite the results. In a visible reply, user TeortaxesTex expressed zero faith in related agentic coding leaderboard claims involving Ox-Alpha, Qwen 3.8 27B, and GPT variants, and urged Reddy to abandon what he called her grift.
Combined views
179.2K
3 posts, first seen 2d ago
Reactions from ranked influencers
3 posts
Ox-Alpha Underperforms And Is Worse Than Last Generation Models Given all the hype we decided to evaluate Ox-Alpha and it turned out to be quite bad Its scores alongside Kimi 2.6 which is 2 generations old However, it's pure marketing genius. Everyone is talking about it
I have zero faith in the idea that ox alpha is < Qwen 3.8 27B on agentic coding nor, for that matter, that Qwen 27B is > GPT 5.6 Sol Max Bindu should burn her grift to the ground if that's what she's getting
What if we rank by Agentic Coding though more sensible, but oh wow, Smaug > Fable https://twitter.com/bindureddy/status/2090949366165152213
@teortaxesTex The big issue with agentic coding on LiveBench is that it's easily benchmaxxed This is the outcome of the bench-maxxing. We are working on replacing this benchmark with LiveBench 2.0 which will be harder to benchmax