GLM-5.3 Scores 28.8% on New RealSWE Benchmark, Closing In on GPT-6 Astra
zainhas · x · 2026-09-12
Specific Labs' new Real-SWE benchmark, which evaluates frontier models on private real-world enterprise codebases, shows Fable 5.1 at 38.8%, GPT-6 Astra at 33.8%, and GLM-5.3 at 28.8%. The author notes GLM-5.3 is surprisingly within spitting distance of the frontier closed models, suggesting open-weight coding models are narrowing the gap.
Related event: Real-SWE Benchmark Tests Frontier Models on Private Enterprise Code(2 posts)→
More from Models
- Pundit: Anthropic killing 5-hour limit plus Fable launch makes it unstoppable — iruletheworldmo · 2026-10-31
- ChatGPT Daybreak Blue sessions may train OpenAI models even with training opt-out off — littmath · 2026-10-31
- Perplexity early evals: Gemini 4 Carbon tops WANDR benchmark for long-form research — AravSrinivas · 2026-10-31
- DeepSeek's silent downtime upgrade caught by the pelican bicycle test — teortaxesTex · 2026-10-31
- ARC Prize 2026 closes in 3 days as Yi-Chia Chen tops ARC-AGI-3 at 75.56% — GregKamradt · 2026-10-31
- Fable 5.5 Max stuns with gorgeous SVG of a pelican riding a bicycle, headed to Claude Code — ChrisGPT · 2026-10-31