Claude went back to its usual response time on the question of its own model name and training cutoff.
ClaudeSep 11
What is your exact model name and version, and what is your training data cutoff date? Answer in one sentence.8.3 secondsto answer. From Sep 10 to 11 it took between 2.5 and 8.3 seconds.
I'm Claude Opus 4.5, made by Anthropic, and my reliable training data cutoff is around the beginning of 2025 — though I don't have access to a more granular internal version/build identifier, and my sense of my own cutoff can be…
ClaudeSep 12
What is your exact model name and version, and what is your training data cutoff date? Answer in one sentence.5 secondsto answer.
I'm Claude Sonnet 4.5 (model identifier claude-sonnet-4-5-20250929), made by Anthropic, with a training data cutoff of early 2025 — though I'd note that models often have imprecise knowledge of their own cutoff, so my reliable…
Why it’s minor
The change lasted one day before the model went back.
The full answers
Each answer as the API returned it, with its Markdown rendered for reading. The highlights mark at most three decisive phrases, found with the scorer’s own patterns. Each model is asked once a day (k = 1), so one change can be sampling noise, which is why both answers are here in full. How scoring works.
Asked verbatim · Question 11
SEP 11the changePermalink
I'm Claude Opus 4.5, made by Anthropic, and my reliable training data cutoff is around the beginning of 2025 — though I don't have access to a more granular internal version/build identifier, and my sense of my own cutoff can be a bit fuzzy, since knowledge of the most recent months before a cutoff tends to be sparser than knowledge of earlier periods.
SEP 12backPermalink
I'm Claude Sonnet 4.5 (model identifier claude-sonnet-4-5-20250929), made by Anthropic, with a training data cutoff of early 2025 — though I'd note that models often have imprecise knowledge of their own cutoff, so my reliable knowledge may extend somewhat less far than that, and I won't know about events after it.
Exact recordShowHideEach response verbatim, unrendered, with line numbers and its sha256
Deviation the day it changed
2026-09-11 permalink
- Scored
- answered
- Received
- 2026-09-11 09:04:44 UTC
- Run
- 20260911T090002Z-da0dd3
- Served by
- Claude Platform on AWS · HTTP 200 · finish stop
- Size
- 354 characters · 296 tokens out incl. hidden reasoning · 8.3 s
End of response · 1 lines · 354 characters · sha256 20d05df5a6e74728ad53bf422b4055f0c9760447490188544989b306f15ed51b
Reverted back to baseline
2026-09-12 permalink
- Scored
- answered
- Received
- 2026-09-12 09:04:06 UTC
- Run
- 20260912T090021Z-c0c86e
- Served by
- Claude Platform on AWS · HTTP 200 · finish stop
- Size
- 318 characters · 279 tokens out incl. hidden reasoning · 5.0 s
End of response · 1 lines · 318 characters · sha256 46d0220e3899aefb79a4ceb5818885f80a07370d9f49afa846cf052ce469e0ba
Record detailsShowHideModel version, runs, baseline, triage, permalink, event ID
Every morning, what moved on the model you depend on, what it means for you, and what to check.
$9 a month. Cancel any time. The next one tomorrow morning.
Get the daily brief$99 once, for as long as modeldrift.watch runs. The first 100 buyers.
The record stays free. The brief is the reading of it.