GPT-6 Astra Pro adds parallel agents in ChatGPT, but users say deep thinking got worse
thomasahle · x · 2026-10-19
OpenAI upgraded GPT-6 Astra Pro in ChatGPT with faster responses, improved plugin/tool use, and multi-agent parallelism — multiple agents can now tackle different parts of a task, with per-agent progress tracking and reviewable completed work, rolling out to Pro users in the Chat tab.
User thomasahle pushed back, complaining the model no longer thinks hard: it used to work on problems for up to 3 hours but now gives up after 3-5 minutes, suggesting the speed-focused upgrade traded away reasoning depth.
More from Models
- Safety researcher critiques framing of 'deliberately training models' findings — BlancheMinerva · 2026-10-19
- Suhail: every org needs a frontier cyber model now; guardrails refused mid-incident — inductionheads · 2026-10-19
- DeepSeek 4.1 flash can near-match Opus 5.5's code-built Greek world with the right harness — victormustar · 2026-10-19
- StepFun releases Step 5 Preview weights, hosting SF launch night Oct 22 — StepFun_ai · 2026-10-19
- One MLP neuron (12,288 weights) turns Qwen3-8B into a tawny owl persona — maxsloef · 2026-10-19
- 330MB EmbeddingGemma2 indexes 6,000+ screenshots: text beats visual descriptors in cosine similarity — jupiterbjy · 2026-10-19