Studio-level card: key art, verdict, results and case-file panels (28 Oct 2026)

.gitattributes

@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+panels/case.webp filter=lfs diff=lfs merge=lfs -text
+panels/results.webp filter=lfs diff=lfs merge=lfs -text
+panels/verdict.webp filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,56 +1,51 @@
----
-pipeline_tag: text-generation
-license: cc-by-4.0
-tags:
-  - glyph-cluster
-  - stealth-model
-  - stealthmark
-  - coding
-  - reasoning
-  - long-context
-  - tool-use
-thumbnail: https://stealthmodels.com/glyph-cluster/assets/social-card-v2.png
----
-
-<table>
-  <tr>
-    <td width="65%" valign="middle">
-      <h1>Glyph Cluster</h1>
-      <p><em>A little company of characters.</em></p>
-      <p><strong>Still in stealth.</strong></p>
-      <p>Glyph Cluster is an unclaimed model built for coding and long-context analysis, with a 256K context and tool calling.</p>
-      <p><a href="https://stealthmodels.com/glyph-cluster/"><strong>Meet Glyph Cluster ↗</strong></a></p>
-    </td>
-    <td width="35%" align="center" valign="middle">
-      <a href="https://stealthmodels.com/glyph-cluster/">
-        <img src="https://stealthmodels.com/glyph-cluster/assets/social-card-v2.png" width="260" alt="Glyph Cluster with four colorful glyph characters pulling into a block puzzle.">
-      </a>
-    </td>
-  </tr>
-</table>
-
-> Independent benchmark card by [Stealth Models](https://stealthmodels.com/): our test results and findings for this model.
-
-## Facts
-| First seen | Revealed | Developer | Context | Input | Tool calls | Speed in our tests | Access |
-|---|---|---|---|---|---|---|---|
-| 7 Oct 2026 | not yet | unknown | 256K tokens | Text only | Yes (function calling) | about 69 tok/s | Active as a free preview; prompts may be retained for training |
-
-## StealthMark results
-| Benchmark score | General Intelligence | Visual Intelligence |
-|---|---|---|
-| 18.9 | 12.0 | 29.8 |
-
-Its highest result is on [Beach](https://stealthmodels.com/benchmarks/models/glyph-cluster/beach), creativity, perspective and visual beauty, at 46.3 out of 100.
-StealthMark 1.8.8, results read 28 October 2026. [Full results](https://stealthmodels.com/benchmarks/models/glyph-cluster) · [On the map](https://stealthmodels.com/benchmarks/map)
-
-<img src="https://stealthmodels.com/benchmarks/media/f72d7dfcd65a0932.png" width="360" alt="Glyph Cluster's Beach result, its highest-scoring benchmark task.">
-
-## Clues so far
-- **9 Oct 2026, our lead:** Microsoft's MAI. Glyph's token counts matched MAI Thinking 1 and our OpenAI GPT reference models on all 24 phrases, which points to the `o200k_base` tokenizer. Our [verdict](https://stealthmodels.com/glyph-cluster/) is that Glyph Cluster appears to be a slight improvement over MAI Thinking 1.
-- **9 Oct 2026, our observation:** the Newton's Mirror results of Glyph and MAI Thinking 1 share the same physical failures, and their Beach results share an unusual choice of pale animal figures.
-- **25 Oct 2026:** Iceman Contributor counts tokens exactly like Glyph Cluster, MAI Thinking 1 and OpenAI's GPT-6 models (23 of 23 phrases), so the same tokenizer family sits behind both stealth models. See the [Iceman card](https://huggingface.co/stealth-model/iceman-contributor).
-
----
-[Model guide](https://stealthmodels.com/glyph-cluster/) · [StealthMark](https://stealthmodels.com/stealthmark) · [Play the benchmarks](https://stealthmodels.com/play) · [Follow for new reveals](https://huggingface.co/stealth-model)
-This card and its images: CC BY 4.0, credit stealthmodels.com. Updated 28 October 2026.
+---
+pipeline_tag: text-generation
+license: cc-by-4.0
+tags:
+  - glyph-cluster
+  - stealth-model
+  - stealthmark
+  - coding
+  - reasoning
+  - long-context
+  - tool-use
+thumbnail: https://stealthmodels.com/glyph-cluster/assets/social-card-v2.png
+---
+
+<a href="https://stealthmodels.com/glyph-cluster/?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card"><img src="https://huggingface.co/stealth-model/glyph-cluster/resolve/main/panels/hero.webp" alt="Glyph Cluster with four colorful glyph characters pulling into a block puzzle, plus 256K text context and seven SVG benchmarks."></a>
+
+**Glyph Cluster's token counts match Microsoft's MAI Thinking 1 on all 24 test phrases, which is our lead for who made it.**
+
+*Independent benchmark card by [Stealth Models](https://stealthmodels.com/?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card): our test results and findings for this model.*
+
+<a href="https://stealthmodels.com/benchmarks/models/glyph-cluster?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card"><img src="https://huggingface.co/stealth-model/glyph-cluster/resolve/main/panels/verdict.webp" alt="StealthMark 1.8.8 results for Glyph Cluster: benchmark score 18.9 out of 100, General Intelligence 12.0, Visual Intelligence 29.8. Strongest at composing a whole scene: 46.3 on Beach."></a>
+
+<a href="https://stealthmodels.com/benchmarks/models/glyph-cluster?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card"><img src="https://huggingface.co/stealth-model/glyph-cluster/resolve/main/panels/results.webp" alt="Every StealthMark result of Glyph Cluster: Beach 46.3, Raccoon 35.7, Selfie 32.4, Ground Truth 19.9, Newton's Mirror 12.8, Mona Lisa 11.7, Foresight 6.4, Latch 3.6."></a>
+
+**Open each result** · Visual Intelligence: [Beach&nbsp;46.3](https://stealthmodels.com/benchmarks/models/glyph-cluster/beach?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [Raccoon&nbsp;35.7](https://stealthmodels.com/benchmarks/models/glyph-cluster/raccoon?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [Selfie&nbsp;32.4](https://stealthmodels.com/benchmarks/models/glyph-cluster/selfie?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [Mona&nbsp;Lisa&nbsp;11.7](https://stealthmodels.com/benchmarks/models/glyph-cluster/mona-lisa?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card)<br>
+General Intelligence: [Ground&nbsp;Truth&nbsp;19.9](https://stealthmodels.com/benchmarks/models/glyph-cluster/ground-truth?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [Newton's&nbsp;Mirror&nbsp;12.8](https://stealthmodels.com/benchmarks/models/glyph-cluster/newtons-mirror?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [Foresight&nbsp;6.4](https://stealthmodels.com/benchmarks/models/glyph-cluster?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card#game-foresight-self-improvement) · [Latch&nbsp;3.6](https://stealthmodels.com/benchmarks/models/glyph-cluster?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card#game-latch-five-levels)
+
+**[All results and comparisons on stealthmodels.com ↗](https://stealthmodels.com/benchmarks/models/glyph-cluster?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card)**
+
+<a href="https://stealthmodels.com/glyph-cluster/?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card"><img src="https://huggingface.co/stealth-model/glyph-cluster/resolve/main/panels/case.webp" alt="The case file. First seen 7 Oct 2026, revealed not yet, developer unknown, context 256K tokens, input Text only, tool calls Yes (function calling), speed in our tests about 69 tok/s. 9 Oct 2026, our finding: Our lead: Microsoft’s MAI. Its token counts and its Newton’s Mirror failures match MAI Thinking 1. 25 Oct 2026, our finding: Iceman Contributor counts tokens exactly like Glyph Cluster: one tokenizer family behind both."></a>
+
+Sources: [our MAI lead and verdict](https://stealthmodels.com/glyph-cluster/?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [the Iceman Contributor card](https://huggingface.co/stealth-model/iceman-contributor)
+
+<details><summary>Facts as text</summary>
+
+| Fact | Value |
+|---|---|
+| First seen | 7 Oct 2026 |
+| Revealed | not yet |
+| Developer | unknown |
+| Context | 256K tokens |
+| Input | Text only |
+| Tool calls | Yes (function calling) |
+| Speed in our tests | about 69 tok/s |
+| Access | Active as a free preview; prompts may be retained for training |
+
+</details>
+
+---
+[Model guide](https://stealthmodels.com/glyph-cluster/?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [StealthMark](https://stealthmodels.com/stealthmark?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [Play the benchmarks](https://stealthmodels.com/play?utm_source=huggingface&utm_medium=referral&utm_campaign=glyph-cluster-card) · [Follow for new reveals](https://huggingface.co/stealth-model)<br>
+This card and its images: CC BY 4.0, credit stealthmodels.com. Updated 28 October 2026.