A/B/A tests show GPT-5.6 and GPT-6 both fail on Instant while Medium works, with HTTP 200 but vanishing replies
echoabenteuer · reddit · 2026-10-13
A long-time ChatGPT power user ran Medium→Instant→Medium tests on both GPT-5.6 and GPT-6 with strikingly consistent results:
- All 4 Medium runs succeeded; both Instant runs failed — GPT-5.6 lost the generated response entirely, GPT-6 ended a story prematurely, leaving explicitly requested events unfinished.
- The author notes the two symptoms differ and may not share a root cause, but the matching A/B/A pattern warrants investigation.
Technical evidence: Earlier packet capture via Chrome DevTools showed HTTP 200 OK on POST /backend-api/f/conversation, a completed SSE stream, a visible response during streaming — then the assistant message disappearing immediately after. The issue reproduced with GPT-5.5 in a fresh conversation and persisted across Windows, iOS, Android, and after clearing cookies.
The investigation is documented under OpenAI Support Case #16904787, with similar reports from other community users. The author suspects Instant's effective behavior has quietly changed, pushing long-form work toward Medium.
Related event: Tests show ChatGPT Instant mode silently drops or truncates responses(2 posts)→
More from Models
- Gemini 4 Argon pricing ($2/$10) spotted on Google's SKU page, global release imminent — gaganghotra_ · 2026-10-13
- A style guide for making Claude-like models write normally, and its linguistic traps — gleech · 2026-10-13
- OpenAI denies Pro 200 grandfathering to a subscriber billed through the eligibility window — IrrelevantNecessity · 2026-10-13
- autotrust's JEV-35B-NVFP4, a Qwen3.5 MoE-based decision model, hits HF trending — autotrust · 2026-10-13
- Theo benchmarks Opus 5.5, GPT-6.1 Sol and Muse Wonder Full on a bug fix — shadcn · 2026-10-13
- Claude Code at $100/month plus $20 ChatGPT: AI subscription costs stack up — dumpshoot · 2026-10-13