Add pinned 28M and 123M GGUF conversions and upstream relations

.gitattributes

@@ -42,3 +42,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 5M/timestral-5m-f16.gguf filter=lfs diff=lfs merge=lfs -text
 5M/timestral-5m-f32.gguf filter=lfs diff=lfs merge=lfs -text
 5M/timestral-5m-q8_0.gguf filter=lfs diff=lfs merge=lfs -text
+123M/timestral-123m-f32.gguf filter=lfs diff=lfs merge=lfs -text
+28M/timestral-28m-f16.gguf filter=lfs diff=lfs merge=lfs -text
+28M/timestral-28m-q8_0.gguf filter=lfs diff=lfs merge=lfs -text
+123M/timestral-123m-f16.gguf filter=lfs diff=lfs merge=lfs -text
+28M/timestral-28m-f32.gguf filter=lfs diff=lfs merge=lfs -text

123M/LICENSE

@@ -0,0 +1,10 @@
+Modified MIT License 
+
+Attribution notice: 2026 - Mistral AI
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of the weights of this model and associated documentation files (the “Model”), to deal in the Model without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Model, and to permit persons to whom the Model is furnished to do so, subject to the following conditions: 
+
+1. The above attribution notice and this permission notice shall be included in all copies or substantial portions of the Model. 
+2. You are not authorized to exercise any rights under this license if the global consolidated monthly revenue of your company (or that of your employer) exceeds $20 million (or its equivalent in another currency) for the preceding month. This restriction in (b) applies to the Model and any derivatives, modifications, or combined works based on it, whether provided by Mistral AI or by a third party. You may contact Mistral AI (sales@mistral.ai) to request a commercial license, which Mistral AI may grant you at its sole discretion, or choose to use the Model on Mistral AI's hosted services available at https://mistral.ai/.
+
+THE MODEL IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL MISTRAL AI BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE MODEL OR THE USE OR OTHER DEALINGS IN THE MODEL.

123M/timestral-123m-f16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf3e7eb0ab9581407d9edc2f4b2d1b8dda8c3a39674dbf3b6967971f906c54b0
+size 246787936

123M/timestral-123m-f32.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90a2cad5bff9ac6ef695689b03d459df1cdfb5bb140755da7b99d01cfa0cbeb4
+size 491523712

28M/LICENSE

@@ -0,0 +1,10 @@
+Modified MIT License 
+
+Attribution notice: 2026 - Mistral AI
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of the weights of this model and associated documentation files (the “Model”), to deal in the Model without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Model, and to permit persons to whom the Model is furnished to do so, subject to the following conditions: 
+
+1. The above attribution notice and this permission notice shall be included in all copies or substantial portions of the Model. 
+2. You are not authorized to exercise any rights under this license if the global consolidated monthly revenue of your company (or that of your employer) exceeds $20 million (or its equivalent in another currency) for the preceding month. This restriction in (b) applies to the Model and any derivatives, modifications, or combined works based on it, whether provided by Mistral AI or by a third party. You may contact Mistral AI (sales@mistral.ai) to request a commercial license, which Mistral AI may grant you at its sole discretion, or choose to use the Model on Mistral AI's hosted services available at https://mistral.ai/.
+
+THE MODEL IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL MISTRAL AI BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE MODEL OR THE USE OR OTHER DEALINGS IN THE MODEL.

28M/timestral-28m-f16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64b5b836c54c2d3a83b428e6b9e44e2bbf668c60297245c97a0df160d2c5366d
+size 57008064

28M/timestral-28m-f32.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbb5a8dcb8b224a45b149dc802428beca20504e71f585d942b0530f67f46a89c
+size 112377536

28M/timestral-28m-q8_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:910e94210a8e87c1b67f8c84643cec5328db7d2d11ef833b5b7a4eab8b5ba028
+size 31053504

README.md

@@ -4,6 +4,8 @@ license_name: mistral-modified-mit
 license_link: LICENSE
 base_model:
 - mistralai/Timestral-5M
+- mistralai/Timestral-28M
+- mistralai/Timestral-123M
 - mistralai/Timestral-209M
 base_model_relation: quantized
 library_name: timestral.cpp
@@ -15,7 +17,7 @@ tags:
 ---
 # Timestral GGUF models for timestral.cpp
 
-This interim release contains nine GGUF files for native CPU forecasting with [timestral.cpp](https://github.com/mudler/timestral.cpp).
+This interim release contains fourteen GGUF files for native CPU forecasting with [timestral.cpp](https://github.com/mudler/timestral.cpp).
 Inference requires no Python or libtorch. Conversion uses the pinned upstream checkpoints listed in `manifest.json`.
 
 ## Available models
@@ -23,10 +25,13 @@ Inference requires no Python or libtorch. Conversion uses the pinned upstream ch
 | Directory | Available formats | Validation |
 | --- | --- | --- |
 | `5M/` | F32, F16, Q8_0 | All three pass storage, schema, lineage, and byte-encoding audits. F32 passes trained parity; reduced-format forecast quality is not yet evaluated. |
+| `28M/` | F32, F16, Q8_0 | Storage, schema, source-byte association, RoPE, and encoding audits pass. Native inference and forecast quality remain pending. |
+| `123M/` | F32, F16 | Storage, schema, source-byte association, RoPE, and encoding audits pass. Native inference and forecast quality remain pending. |
 | `209M/` | F32, F16, Q8_0, Q6_K, Q5_K, Q4_K | All six pass cohort/storage admission. Full quality evaluation completes, but the Q8 mean-deviation gate fails. |
 
-**28M and 123M GGUFs are not included.** Their upstream checkpoints are acquired; conversion and native qualification remain pending.
-The intended series includes all four models. This release does not claim that scope is complete.
+All four upstream models now have converted artifacts in this repository. Native qualification is incomplete for 28M and 123M.
+123M Q8_0 is unavailable: its 1712-wide FFN matrices do not satisfy the unchanged 32-value block requirement.
+K formats for 5M, 28M, and 123M are incompatible with the unchanged complete weight policy. No padding or selective exclusions are used.
 
 ## Quality and performance limitations
 
@@ -71,6 +76,6 @@ Keep the original notices with copies and derivatives. This repository does not
 ## Integrity and provenance
 
 `manifest.json` records exact sizes, SHA-256 hashes, upstream revisions, and runtime/reference identities.
-`SHA256SUMS` contains checksums for all nine GGUF files.
+`SHA256SUMS` contains checksums for all fourteen GGUF files.
 F32/F16/Q8_0 use the pinned converter; 209M K formats use the reviewed native quantizer.
 The 5M F32 file is byte-identical to the previously accepted trained-parity artifact.

SHA256SUMS

@@ -7,3 +7,8 @@ bf7605545245a2dd4d32f7ac2f7bbe5c2a12d8f47370814a254862150255b8e8  209M/timestral
 49f364c750906c2a0513a407af5c94cf16a6d711976ce803ef966b17a0f6fe4d  209M/timestral-209m-q6_k.gguf
 e894e488baf17ba07a8dc20c685506a5a7bb89ca8ddb82e920cb2113401d1fef  209M/timestral-209m-q5_k.gguf
 05b9a0c99c099ba23ecb0a8dabf1552a6f0fbd0b7e37df0163fe41c851d575a6  209M/timestral-209m-q4_k.gguf
+90a2cad5bff9ac6ef695689b03d459df1cdfb5bb140755da7b99d01cfa0cbeb4  123M/timestral-123m-f32.gguf
+64b5b836c54c2d3a83b428e6b9e44e2bbf668c60297245c97a0df160d2c5366d  28M/timestral-28m-f16.gguf
+910e94210a8e87c1b67f8c84643cec5328db7d2d11ef833b5b7a4eab8b5ba028  28M/timestral-28m-q8_0.gguf
+cf3e7eb0ab9581407d9edc2f4b2d1b8dda8c3a39674dbf3b6967971f906c54b0  123M/timestral-123m-f16.gguf
+fbb5a8dcb8b224a45b149dc802428beca20504e71f585d942b0530f67f46a89c  28M/timestral-28m-f32.gguf

manifest.json

@@ -67,6 +67,41 @@
       "sha256": "05b9a0c99c099ba23ecb0a8dabf1552a6f0fbd0b7e37df0163fe41c851d575a6",
       "upstream": "mistralai/Timestral-209M",
       "revision": "9874d252ab0aae16f7f7c999f98f98d6ad7d8365"
+    },
+    {
+      "path": "123M/timestral-123m-f32.gguf",
+      "bytes": 491523712,
+      "sha256": "90a2cad5bff9ac6ef695689b03d459df1cdfb5bb140755da7b99d01cfa0cbeb4",
+      "upstream": "mistralai/Timestral-123M",
+      "revision": "5a147287addda85f6acd8da3fb8858fe9af7e8f7"
+    },
+    {
+      "path": "28M/timestral-28m-f16.gguf",
+      "bytes": 57008064,
+      "sha256": "64b5b836c54c2d3a83b428e6b9e44e2bbf668c60297245c97a0df160d2c5366d",
+      "upstream": "mistralai/Timestral-28M",
+      "revision": "49b26409ac87b5935539a5d8ffe7f397b720db47"
+    },
+    {
+      "path": "28M/timestral-28m-q8_0.gguf",
+      "bytes": 31053504,
+      "sha256": "910e94210a8e87c1b67f8c84643cec5328db7d2d11ef833b5b7a4eab8b5ba028",
+      "upstream": "mistralai/Timestral-28M",
+      "revision": "49b26409ac87b5935539a5d8ffe7f397b720db47"
+    },
+    {
+      "path": "123M/timestral-123m-f16.gguf",
+      "bytes": 246787936,
+      "sha256": "cf3e7eb0ab9581407d9edc2f4b2d1b8dda8c3a39674dbf3b6967971f906c54b0",
+      "upstream": "mistralai/Timestral-123M",
+      "revision": "5a147287addda85f6acd8da3fb8858fe9af7e8f7"
+    },
+    {
+      "path": "28M/timestral-28m-f32.gguf",
+      "bytes": 112377536,
+      "sha256": "fbb5a8dcb8b224a45b149dc802428beca20504e71f585d942b0530f67f46a89c",
+      "upstream": "mistralai/Timestral-28M",
+      "revision": "49b26409ac87b5935539a5d8ffe7f397b720db47"
     }
   ]
 }