Tracker›Changelog›2026-08-23-model-gpt-5-6-zu-gpt-5-6-thinking

Model change gpt-5-6 → gpt-5-6-thinking on 2 prompts

observedstructure · model2026-08-23-model-gpt-5-6-zu-gpt-5-6-thinking

Instrument state: Schema 14 · v2 · Bookmarklet 2.2 · measure · gpt-5-6

Evidence

Before
gpt-5-6
After
gpt-5-6-thinking
Breadth
2/31 prompts · local
Modes
instant
Measuring days
23 Aug 2026 · 2 field diffs in events.jsonl
Raw data excerpt (K4-01__instant, 2026-08-23)
{
  "routing.model_slug": "gpt-5-5"
}

Daily data of 23 Aug 2026 →

Context

model/gpt-5-6-thinking model/gpt-5-6

Impact

For tools: model_slug in the send payload and in routing.*. Standard mode gpt-5-6, thinking gpt-5-6-thinking. In the UI „GPT-5.6 Sol". A change of the slug is a baseline break of the time series (see version history).

For visibility: Model generations differ in when they search, how they phrase queries and which format web.run uses. A model change can shift visibility without anything changing on a website.

Status history

23 Aug 2026observed

Open

On 21 Aug two prompts still answered with gpt-5-5 (K4-02, K8-01), from 22 Aug consistently gpt-5-6. Rollout in waves or chance?

How do gpt-5-6 and gpt-5-6-thinking differ in fan-out? Observation: thinking bundles more queries on product questions (up to 7) but barely thinks longer (2–6 s).