Ox Alpha scores 63% on DeepSWE benchmark with mixed user feedback
AccBalanced · x · 2026-08-23
Developer @davis7 ran the DeepSWE benchmark on the mysterious Ox Alpha model, resulting in a 63% score, comparable to DeepSeek V4 Pro and Grok 4.6. Hands-on reviews highlight its superior "voice" and design over Claude/GPT, along with strong handling of sub-agents and complex coding tasks. However, users noted issues like leftover dead code and significant latency at higher reasoning levels. If it is a small open-weight model (like GLM) with low compute needs, it offers insane value.
Related event: Ox Alpha scores ~63% on full DeepSWE; earlier 80% was subset artifact(6 posts)→
More from Models
- Unredacted Anthropic filing exposes internal data-quality debate months before Opus 3 cutoff — zacharynado · 2026-10-15
- Google says teams across the company are using Gemini 4 Argon on its hardest engineering problems — CurieuxExplorer · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- 797GB of n-gram counts in CPU RAM let 1B models match 57% more compute — PangWeiKoh · 2026-10-15
- Leak: Moonshot's Kimi flagship 'x4' done with pre-training, K3.1 in internal testing — lyraxana · 2026-10-15
- Unsloth ships GGUF quants for Nemotron 3.5: Bolt-74B runs in 24GB RAM — danielhanchen · 2026-10-15