The Latent Score: 2026-09-06
Release tli-2026-v4.2-2026-09-06. Methodology tli-2026-v4.2.
As of 2026-09-06, Fable 5.1 leads The Latent Score on the point estimate at 128.2 points among 19 ranked models. Rankings describe published capability evidence, not a statistically proven winner or the best model for every workload.
This page preserves a specific publication. Its data does not change when a new release is published.
Published rankings
| Rank | Model and evidence | Score (points) | Stability range (points) | Contributing results |
|---|---|---|---|---|
| 1 | Fable 5.1 | 128.2 | 122.7–134.2 | 57 |
| 2 | GPT-6 Astra | 124.4 | 114.1–150 | 44 |
| 3 | Opus 5 | 121.4 | 115–126.2 | 84 |
| 4 | Fable 5 | 118.1 | 112.9–122.6 | 96 |
| 5 | Muse Spark 1.3 | 111.9 | 105.5–118.4 | 22 |
| 6 | GPT-5.6 Sol | 108.1 | 103.7–114.4 | 103 |
| 7 | Kimi K3 | 104.7 | 98.6–109.3 | 87 |
| 8 | GLM-5.3 | 103.2 | 95.4–108.4 | 55 |
| 9 | Gemini 3.8 Flash | 102.5 | 94–110.4 | 50 |
| 10 | Muse Spark 1.2 | 102.2 | 93.1–107.3 | 49 |
| 11 | Grok 4.6 | 102.0 | 97–108.5 | 57 |
| 12 | Qwen3.8-Max | 97.1 | 89.7–101.5 | 71 |
| 13 | Gemini 3.7 Flash | 95.0 | 86.4–105 | 60 |
| 14 | GPT-5.6 Terra | 94.5 | 90.2–102.6 | 74 |
| 15 | DeepSeek V4 Pro | 89.4 | 82.6–97.5 | 53 |
| 16 | Sonnet 5 | 88.5 | 82.1–93.9 | 67 |
| 17 | GPT-5.6 Luna | 86.0 | 81.6–94.5 | 71 |
| 18 | DeepSeek V4 Flash | 72.9 | 67.7–82.8 | 53 |
| 19 | Qwen3.8-27B | 69.5 | 59.4–79.9 | 51 |
Methodology and limitations in this release
The scale is anchored to mean 100, SD 15 across 18 frozen reference models. It is not human IQ or a percentage. 104 benchmarks contribute across 7 domains.
This older release did not preserve a written methodology in its public snapshot. Its recorded scores and scale are archived here; current methodology documentation must not be assumed to describe this historical calculation.
Reading this release
Compare model scores alongside their stability ranges and domain coverage. A higher point estimate does not establish superiority for every task. The linked model records disclose contributing and excluded results and their published configurations.
This release uses methodology version tli-2026-v4.2. Comparisons with another release should check both its methodology version and evidence date; a new data release does not necessarily mean the scoring method changed.
Citation and downloads
The Latent. “The Latent Score.” Release tli-2026-v4.2-2026-09-06, dataset dated 2026-09-06. Permanent release page.
Download public scores with definitions (JSON) · Data usage terms