Astra saturates benchmarks but disappoints in real use, pushing developers to Anthropic
torchcompiled · x · 2026-11-18
Developer knowclarified (quoted) complains that GPT 6.2 Astra saturated most benchmarks yet tested demonstrably worse than Fable and even Opus 4.5 in his own use, with Astra 6 feeling worse than 6.1 — so he cancelled his Codex subscription and went all-in on Anthropic, though he'll give the new 6.2 models a shot.
torchcompiled adds a possible explanation: the model may be genuinely strong but extremely particular about prompts or overly specialized in frontier problems — if it doesn't generalize to common use cases and has poor UX, a tool is only as good as you can reasonably control it.
More from Models
- Google ships Gemini 4 Hydrogen, a faster robotics model, plus Antigravity Robotics plugin — keerthanpg · 2026-11-18
- Epoch: GPT-6.2 Astra scores 174 on ECI, above the pre-Astra reasoning-era trendline — airesearch12 · 2026-11-18
- Claude delegates synthetic data generation to local Qwen3.8-27B, citing its own guardrails — maksym_andr · 2026-11-18
- GPT-6.2 stuck in compaction loop in Codex, burning 4.44M tokens in 9 minutes — Dear_Lock_5280 · 2026-11-18
- Inkling-2-Small runs on two DGX Spark units and is claimed competitive with DSv4.1 and GLM 5.3 — LiTianleli · 2026-11-18
- Users find GPT-6.2 keeps setting its generated games in the same fictional place — kieranklaassen · 2026-11-18