michalharakal · Oct 12, 2026 · Oct 12, 2026 · Oct 12, 2026 · Oct 12, 2026 · Oct 12, 2026
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,6 +2,17 @@
 
 ## [Unreleased]
 
+## [0.58.0] - 2026-10-12
+
+Headline: **input front ends and model identity (SKEEP-006).** Tokenizers leave `skainet-io-core`
+for three new modules with one front-end contract across modalities, families become string
+identities with descriptors in an immutable registry, and `tokenizer.json` support grows to
+Metaspace BPE, Unigram and WordPiece with a golden corpus recorded against the Hugging Face
+`tokenizers` library. Every 0.57 call site keeps compiling through deprecated aliases.
+
+Version note: SKaiNET-transformers 0.58.0 follows on this engine release and resolves every
+tokenizer through the engine; the two stay on one version line.
+
 ### Added
 
 - **Input front ends and model identity (SKEEP-006, draft).** Three new modules:
@@ -27,8 +38,8 @@
   the per-format `CheckpointView` implementations the resolver consumes.
 
 ### Fixed
-- `tokenizer.json` added tokens keep their `lstrip` / `rstrip` / `single_word` flags through `SpecialTokenSplitter`, so the whitespace before an `lstrip` token such as the ModernBERT / mmBERT `<mask>` belongs to the special token as in the reference library (SKEEP-006, #1324 corpus).
 
+- `tokenizer.json` added tokens keep their `lstrip` / `rstrip` / `single_word` flags through `SpecialTokenSplitter`, so the whitespace before an `lstrip` token such as the ModernBERT / mmBERT `<mask>` belongs to the special token as in the reference library (SKEEP-006, #1324 corpus).
 - **`tokenizer.json` parity for byte-level BPE** (#1321): the array form of `model.merges`
   written by every `tokenizers` release since 0.20 loads (#1323); every `added_tokens` entry is
   matched atomically and counted in `vocabSize`, special or not (#1324); the `normalizer` block

diff --git a/CITATION.cff b/CITATION.cff
@@ -12,8 +12,8 @@ authors:
     orcid: https://orcid.org/0009-0006-7981-4075
   - name: SKaiNET-developers contributors
     website: https://github.com/SKaiNET-developers
-version: "0.57.0"
-date-released: "2026-09-24"
+version: "0.58.0"
+date-released: "2026-10-12"
 license: MIT
 repository-code: https://github.com/SKaiNET-developers/SKaiNET
 url: https://skainet-developers.github.io/SKaiNET/

diff --git a/README.md b/README.md
@@ -51,7 +51,7 @@ Add the core dependencies (Gradle Kotlin DSL):
 ```kotlin
 dependencies {
     // Recommended: import the umbrella BOM and drop versions on the engine modules.
-    implementation(platform("sk.ainet:skainet-bom:0.57.0"))
+    implementation(platform("sk.ainet:skainet-bom:0.58.0"))
 
     implementation("sk.ainet.core:skainet-lang-core")
     implementation("sk.ainet.core:skainet-backend-cpu")
@@ -308,18 +308,21 @@ val withoutLabel = dataPipeline<RawDataset>()
 
 ---
 
-## What's New in 0.57.0
+## What's New in 0.58.0
 
-Grouped-query KV-cache graphs export with an attention mask, and the toolchain moves to Kotlin 2.4.20:
+Input front ends and model identity land as SKEEP-006:
 
-- **Masked grouped-query attention exports to valid StableHLO** — a per-head mask under GQA is split
-  into the K/V head groups instead of broadcast onto them, and a mask onto a dynamic key length (the
-  chunked-prefill shape of a KV-cache graph) uses `dynamic_broadcast_in_dim` with the expansion hints
-  IREE 3.11 needs. Static graphs emit exactly what they did before.
-- **kotlinx-io is part of `skainet-data-source`'s API** — callers of `openSource()` / `copyTo()` no
-  longer need to declare it themselves.
-- **Kotlin 2.4.20**, AGP 9.4.1, Ktor client 3.6.0; the `webpack` security pin is retired now that
-  the Kotlin toolchain's own webpack is past the fix.
+- **One front-end contract for every input modality** — `skainet-frontend-core` defines
+  `FrontEnd`, `FrontEndSpec` and the text, audio, image and multimodal contracts; a resolver turns a
+  checkpoint view into the front ends a model needs. The text lane ships in `skainet-frontend-text`.
+- **Tokenizers move out of io-core and cover more files** — Metaspace BPE with byte fallback,
+  Viterbi Unigram and WordPiece join byte-level BPE, SentencePiece and Tekken. A golden corpus
+  recorded with the Hugging Face `tokenizers` library guards exact id parity for GPT-2, Qwen2.5,
+  TinyLlama, T5, BERT, Gemma 4 and the two public Laya tokenizers.
+- **Families are identities** — `skainet-model-identity` carries `ModelIdentity`, `CheckpointView`
+  and an immutable `ModelFamilyRegistry`; the old enum-style detection is deprecated.
+- Everything that compiled against 0.57 still compiles: the io-core names are deprecated aliases
+  that forward to the new homes.
 
 See [CHANGELOG.md](CHANGELOG.md) for full release notes, including every prior release.
 
@@ -345,6 +348,12 @@ We love contributions! Whether it's a new operator, documentation, or a bug fix:
 
 Browse the full codebase documentation on [DeepWiki](https://deepwiki.com/SKaiNET-developers/SKaiNET).
 
+### Contributors (0.58.0)
+
+- **Michal Harakal** ([@michalharakal](https://github.com/michalharakal)) — SKEEP-006: the
+  front-end and model-identity modules, the tokenizer move with Metaspace BPE, Unigram and
+  WordPiece, the golden corpus, and the release
+
 ### Contributors (0.57.0)
 
 - **Michal Harakal** ([@michalharakal](https://github.com/michalharakal)) — masked grouped-query
@@ -428,7 +437,7 @@ from it). Ready-to-paste BibTeX:
   author       = {Harakal, Michal and {SKaiNET-developers contributors}},
   title        = {{SKaiNET}: A {Kotlin} Multiplatform {AI} Framework},
   year         = {2026},
-  note         = {Version 0.57.0. MIT License},
+  note         = {Version 0.58.0. MIT License},
   howpublished = {\url{https://github.com/SKaiNET-developers/SKaiNET}}
 }
 ```

diff --git a/docs/antora.yml b/docs/antora.yml
@@ -15,7 +15,7 @@ asciidoc:
     framework_name: SKaiNET
     # Current SKaiNET release — bump once per release; referenced as
     # {skainet_version} in dependency snippets (blocks need subs="attributes+").
-    skainet_version: 0.57.0
+    skainet_version: 0.58.0
     ksp_version: 2.2.21-2.0.5
     dokka_version: 2.1.0
     asciidoctorj_version: 3.0.0
diff --git a/docs/modules/ROOT/pages/reference/kernel-support-matrix.adoc b/docs/modules/ROOT/pages/reference/kernel-support-matrix.adoc
@@ -1,7 +1,7 @@
 = Kernel × platform support matrix
 :description: Which compute-kernel provider serves each weight format on each KMP target.
 
-Generated from `kernel-support.json` (version `0.57.0`) by `KernelSupportMatrixTest` — registry introspection of the registered `KernelProvider` implementations. Do not edit by hand; run `./gradlew generateKernelMatrix` to refresh.
+Generated from `kernel-support.json` (version `0.58.0`) by `KernelSupportMatrixTest` — registry introspection of the registered `KernelProvider` implementations. Do not edit by hand; run `./gradlew generateKernelMatrix` to refresh.
 
 Each cell is the best (highest-priority) provider that serves `Float32 × format` `matmul` on that platform: *native-ffm* (100) → *panama-vector* (50) → *scalar* (0). An empty cell (`—`) means no provider carries a kernel there (the format is dequant-to-FP32 only).
 

diff --git a/docs/modules/ROOT/pages/reference/operators/generated/index.adoc b/docs/modules/ROOT/pages/reference/operators/generated/index.adoc
@@ -1,6 +1,6 @@
 = AI-NET Operators Reference
 
-Generated from version `0.57.0` on 2026-09-24
+Generated from version `0.58.0` on 2026-10-12
 
 == Operators by Modality
 

diff --git a/docs/modules/ROOT/pages/reference/ops-status-matrix.adoc b/docs/modules/ROOT/pages/reference/ops-status-matrix.adoc
@@ -1,7 +1,7 @@
 = Operator Coverage Matrix
 :description: Cross-backend status for every operator function in SKaiNET.
 
-Generated from `operators.json` version `0.57.0` on 2026-09-24.
+Generated from `operators.json` version `0.58.0` on 2026-10-12.
 
 Rows are `Operator.function` pairs. The `Validated` column shows whether the function's documentation has been DARC-validated by a reviewer (see xref:contributing/darc-workflow.adoc[DARC workflow]). Remaining columns are backends that appear in any function's `statusByBackend` map — a missing entry means the backend makes no claim about the function (treat it as "unknown", not "not supported").
 

diff --git a/docs/modules/ROOT/pages/tutorials/android-classifier-getting-started.adoc b/docs/modules/ROOT/pages/tutorials/android-classifier-getting-started.adoc
@@ -19,12 +19,12 @@ An Android project with Kotlin. Add the SKaiNET modules:
 [source,kotlin]
 ----
 dependencies {
-    implementation("sk.ainet.core:skainet-lang-core:0.57.0")     // tensors, DSL, training
-    implementation("sk.ainet.core:skainet-backend-cpu:0.57.0")   // CPU ops
-    implementation("sk.ainet.core:skainet-compile-dag:0.57.0")   // autograd (training context)
-    implementation("sk.ainet.core:skainet-data-api:0.57.0")      // Dataset / DataBatch
-    implementation("sk.ainet.core:skainet-data-simple:0.57.0")   // embedded Iris
-    runtimeOnly("sk.ainet.core:skainet-backend-jni-cpu:0.57.0")  // NEON kernels (see below)
+    implementation("sk.ainet.core:skainet-lang-core:0.58.0")     // tensors, DSL, training
+    implementation("sk.ainet.core:skainet-backend-cpu:0.58.0")   // CPU ops
+    implementation("sk.ainet.core:skainet-compile-dag:0.58.0")   // autograd (training context)
+    implementation("sk.ainet.core:skainet-data-api:0.58.0")      // Dataset / DataBatch
+    implementation("sk.ainet.core:skainet-data-simple:0.58.0")   // embedded Iris
+    runtimeOnly("sk.ainet.core:skainet-backend-jni-cpu:0.58.0")  // NEON kernels (see below)
 }
 ----
 

diff --git a/gradle.properties b/gradle.properties
@@ -1,5 +1,5 @@
 GROUP=sk.ainet.core
-VERSION_NAME=0.58.0-SNAPSHOT
+VERSION_NAME=0.58.0
 POM_DESCRIPTION=SKaiNET
 
 POM_URL=https://github.com/SKaiNET-developers/skainet/