SIGHT.
Superintelligence Information, Granularity & Transparency

SI models are up to 48% smarter at 3am. Six claims, opened.

Every figure on this page is a claim about a base. Here the claims sit in one column, each carrying its number, each opening onto its evidence, its arithmetic and its source. Collapsed is a reading order, not a withholding: nothing is off the page, and the receipt below never closes.

Endpoint Accuracy Index v1.1 · 8 Oct–4 Nov 2026 Figures current as of 17 November 2026

The claim index

Six lines. Each one is a disclosure.

The receipt · always open

Every figure, its base, its sample, its date.

A percentage is a claim about a base. Every row names the base it divides by, the sample it divides, and the date the source published it. The six lines above point here; this table points at nothing else.

Every figure on this page, with its arithmetic, its base, its sample, and the date and source it comes from.
figurearithmeticbasesampledate · source
48%(99.3 − 67.2) ÷ 67.2 = 47.8%worst endpoint26 endpoints, one model8 Oct–4 Nov 2026 · Artificial Analysis
±6.1 pts√(4.3² + 4.3²) = 6.1gap interval (AA avg ±4.3)26 endpointssame · 39–57% relative
32.1 pts99.3 − 67.20–100 indexsame 26same
10 of 26count, 95% CI wholly < 100reference = 10026 endpointssame
28%(64 − 50) ÷ 50 = 28.0%peak day (50)50 fixed tasks, n=5022–23 May 2026 · Margin Lab
16%0.8% → 16% at worst hourSonnet 4 requests31 Aug 202517 Sep 2025 · Anthropic
5.2%(84·99.3 + 16·67.2)/100 = 94.16; (99.3 − 94.16) ÷ 99.3 = 5.2%best endpoint16% draw the worstderived · 17 Nov 2026
7.6%(58.1 − 54.0) ÷ 54.0 = 7.6%peak base (54.0)24 days, 1,175 graded25 Jan–20 Feb 2026 · Margin Lab Codex
88%(63.6 − 7.6) ÷ 63.6 = 88.1%uncapped endpointHLE-250, 16,384-token cap8 Oct–4 Nov 2026 · Artificial Analysis
4.0–47.8%per-model relative range, median 22%each model’s worstsix models, 103 endpoints8 Oct–4 Nov 2026 · Artificial Analysis

Figures current as of 17 November 2026. Rows are ordered as the claims are; nothing in this table is derived except the two marked derived.

Out of the index

Two sources the six claims do not spend.

The file carries two more measurements. They are evidence of the same shape — a tail where a fixed thing stops being fixed — and they are not part of the six figures above, so they are named here rather than folded in.

vLLM · 22 Apr 2026

“The State of FP8 KV-Cache and Attention Quantization.” 128k needle-in-a-haystack: BF16 91%, FP8 before the accumulation fix 13%, after 89%. The tail is real; the fix is real.

Chen et al. · 2023

Chen, Zaharia & Zou, “How Is ChatGPT’s Behavior Changing over Time?” GPT-4 84.0% → 51.1% on a balanced 1,000-item prime/composite set, March–June 2023 (revised version). Temporal drift, not a clock effect.

The limit

The one measurement nobody has run.

Not a controlled clock experiment. Nobody has published one model, one task set, one endpoint, at 03:00 and 15:00, with the serving endpoint logged. The 48% is the endpoint spread; the 16% is the load-hour exposure; the 5.2% is the expected loss if 16% of requests draw the worst endpoint. The clock experiment is the one measurement the industry still owes. When a provider runs it, this page gets a better number.

Not every model. Six models are in the index; the headline uses the widest, GLM-5.3. Not intelligence. Task success on three evaluations. Not a promise. A measured spread and a routing mechanism.

An index makes the shape of the argument visible: six claims, six figures, one base each. It does not make the argument bigger than its evidence, and the limit above is part of the index too.

Sources

Dated, and checkable.

01

Artificial Analysis, “Endpoint Accuracy Index v1.1.” 26 GLM-5.3 endpoints, 67.2–99.3; 10 below reference; HLE-250 7.6 vs 63.6 under a 16,384-token cap; six models, 103 endpoints, 24 below. Measured 8 Oct–4 Nov 2026. artificialanalysis.aipublished 9 Nov 2026 · fetched 17 Nov 2026

02

OpenRouter, “Provider Routing.” Routing “learns how much load each provider can take before it returns 429s, and sends the rest to other providers.” openrouter.ai/docsfetched 17 Nov 2026

03

Anthropic, “A postmortem of three recent issues.” “We never reduce model quality due to demand, time of day, or server load”; “At the worst impacted hour on August 31, 16% of Sonnet 4 requests were affected,” up from 0.8%, after “a routine load balancing change”; ~30% of Claude Code users hit a misrouted server. anthropic.com17 Sep 2025

04

Anthropic, “Model IDs and versions.” “Model weights are fixed for a given ID, but the serving infrastructure around the model can change over time. This infrastructure includes components such as the request router, safety classifiers, and sampling logic.” platform.claude.comfetched 17 Nov 2026

05

Margin Lab, “Claude Code tracker” and “Methodology.” 50 fixed SWE-Bench-Pro tasks, one run per day at 06:00 UTC, Wilson 95% intervals; the largest same-model daily move is 64% → 50% on 22–23 May 2026 — 14 points, just over its ±13-point daily bar, and flagged by its same-task check at p < 0.001. The corrected 58.1 → 54.0 window is the Codex tracker, 25 Jan–20 Feb 2026. marginlab.aifetched 17 Nov 2026

06

vLLM, “The State of FP8 KV-Cache and Attention Quantization.” 128k needle-in-a-haystack: BF16 91%, FP8 before the accumulation fix 13%, after 89%. The tail is real; the fix is real. vllm.ai22 Apr 2026

07

Chen, Zaharia & Zou, “How Is ChatGPT’s Behavior Changing over Time?” GPT-4 84.0% → 51.1% on a balanced 1,000-item prime/composite set, March–June 2023 (revised version). Temporal drift, not a clock effect. arXiv:2307.090092023

Figures current as of 17 November 2026.

The receipt · 03:00

7.6% was an average.48% is the spread.88% is the edge.

Say what you serve