diff --git a/.github/ci/benchmark_config.json b/.github/ci/benchmark_config.json
index 43f8895d..43677858 100644
--- a/.github/ci/benchmark_config.json
+++ b/.github/ci/benchmark_config.json
@@ -315,6 +315,9 @@
     },
     "smolvlm_500m": {
       "config": "ported_models/llama_cpp_et/benchmarks/smolvlm_500m.json"
+    },
+    "olmoe_1b_7b_0125": {
+      "config": "ported_models/llama_cpp_et/benchmarks/olmoe_1b_7b_0125.json"
     }
   }
 }
diff --git a/ported_models/llama_cpp_et/artifacts.json b/ported_models/llama_cpp_et/artifacts.json
index 9c419ae1..5fe05c10 100644
--- a/ported_models/llama_cpp_et/artifacts.json
+++ b/ported_models/llama_cpp_et/artifacts.json
@@ -485,6 +485,24 @@
       "sha256": "d1eb8b6b23979205fdf63703ed10f788131a3f812c7b1f72e0119d5d81295150",
       "size_bytes": 108783360,
       "note": "SmolVLM 500M vision projector (SigLIP ~93M + MLP). Q8_0 quantized. Must be loaded alongside smolvlm_500m_q8_gguf."
+    },
+    "olmoe_1b_7b_0125_q8_gguf": {
+      "kind": "model",
+      "framework": "llama.cpp-et",
+      "variant": "OLMoE-1B-7B-0125-Q8_0",
+      "filename": "OLMoE-1B-7B-0125-Q8_0.gguf",
+      "env": "OLMOE_1B_7B_0125_MODEL_PATH",
+      "source": {
+        "type": "huggingface",
+        "repo": "allenai/OLMoE-1B-7B-0125-GGUF",
+        "revision": "b8453f3c06477f3efc002ff89f2c3c149ffbeec1",
+        "filename": "OLMoE-1B-7B-0125-Q8_0.gguf",
+        "url": "https://huggingface.co/allenai/OLMoE-1B-7B-0125-GGUF/resolve/b8453f3c06477f3efc002ff89f2c3c149ffbeec1/OLMoE-1B-7B-0125-Q8_0.gguf"
+      },
+      "sha256": "5a82c080fb821ed5afac78768377260f984e19decf3fc49befafe61d7a9d44ad",
+      "size_bytes": 7359942720,
+      "local_cache": "local-artifacts/models/OLMoE-1B-7B-0125-Q8_0.gguf",
+      "board_path": "/data/models/OLMoE-1B-7B-0125-Q8_0.gguf"
     }
   }
 }
diff --git a/ported_models/llama_cpp_et/benchmarks/olmoe_1b_7b_0125.json b/ported_models/llama_cpp_et/benchmarks/olmoe_1b_7b_0125.json
new file mode 100644
index 00000000..d69f1622
--- /dev/null
+++ b/ported_models/llama_cpp_et/benchmarks/olmoe_1b_7b_0125.json
@@ -0,0 +1,52 @@
+{
+  "runner": "llama_server",
+  "board": true,
+  "framework": {
+    "name": "llama.cpp-et",
+    "runner": "llama_server",
+    "source_artifact": "llama_cpp_source"
+  },
+  "artifacts_file": "../artifacts.json",
+  "canonical_variant": "OLMoE-1B-7B-0125-Q8_0",
+  "score": {
+    "metric": "tokens_per_second",
+    "label": "Decode tokens/s",
+    "higher_is_better": true
+  },
+  "llama_server": {
+    "source_artifact": "llama_cpp_source",
+    "model_artifact": "olmoe_1b_7b_0125_q8_gguf",
+    "server_artifact": "llama_server",
+    "workdir_artifact": "llama_cpp_build",
+    "host": "127.0.0.1",
+    "port": 18145,
+    "device": "ET",
+    "gpu_layers": 99,
+    "ctx_size": 2048,
+    "batch_size": 256,
+    "ubatch_size": 128,
+    "parallel": 1,
+    "cache_ram_mib": 0,
+    "ready_timeout_s": 300,
+    "request_timeout_s": 420,
+    "flash_attn": false,
+    "api": "completion",
+    "prompt": "Repeat this token sequence without commentary: OK OK OK OK OK OK OK OK OK OK",
+    "max_tokens": 96,
+    "temperature": 0,
+    "ignore_eos": true,
+    "min_completion_tokens": 32,
+    "perplexity": {
+      "enabled": true,
+      "perplexity_artifact": "llama_perplexity",
+      "corpus_artifact": "wikitext2_raw_test",
+      "ctx_size": 128,
+      "batch_size": 128,
+      "ubatch_size": 128,
+      "timeout_s": 420,
+      "min_ppl": 1.0,
+      "max_ppl": 1000.0,
+      "chunks": 4
+    }
+  }
+}
diff --git a/ported_models/olmoe_1b_7b_0125/docs/RECIPE.md b/ported_models/olmoe_1b_7b_0125/docs/RECIPE.md
new file mode 100644
index 00000000..ff329306
--- /dev/null
+++ b/ported_models/olmoe_1b_7b_0125/docs/RECIPE.md
@@ -0,0 +1,67 @@
+# OLMoE-1B-7B-0125 Porting Recipe
+
+## Overview
+
+Adds `allenai/OLMoE-1B-7B-0125` (7B-total/1B-active-parameter Mixture-of-Experts
+causal LM, 64 experts, 8 active per token) to the `llama_cpp_et` benchmark
+suite. This introduces the **OLMoE** execution family to the board — a
+fully open-source MoE release from Ai2 (open weights, open training data,
+open code).
+
+## Model Reference
+
+- **Source**: `allenai/OLMoE-1B-7B-0125` (Hugging Face), revision
+  `9b0c1aa87e34a20052389dce1f0cf01da783f654`
+- **License**: Apache 2.0
+- **Architecture**: `arch = olmoe` (`OlmoeForCausalLM`), 16 layers, 64
+  experts, 8 experts active per token (confirmed against `config.json`).
+
+## Hosting
+
+Ai2 publishes an official first-party GGUF repository with a full
+quantization ladder, so this uses that directly rather than a self-converted
+artifact:
+
+- **GGUF source**: `allenai/OLMoE-1B-7B-0125-GGUF`, revision
+  `b8453f3c06477f3efc002ff89f2c3c149ffbeec1`, file
+  `OLMoE-1B-7B-0125-Q8_0.gguf`, 7,359,942,720 bytes,
+  `sha256=5a82c080fb821ed5afac78768377260f984e19decf3fc49befafe61d7a9d44ad`
+  (verified against the file's own `x-linked-etag` header, matching the
+  real remote content-length exactly).
+
+## Local Verification (confirmed live, not speculative)
+
+Independently self-converted the base checkpoint (`convert_hf_to_gguf.py
+--outtype q8_0`, no fixes needed, clean conversion first attempt) to
+cross-check the official GGUF is architecturally sound, then ran real
+inference via the CPU-backend build (see verification-tier note in the
+`jamba_tiny_dev` recipe for why CPU, not full ET sysemu):
+
+- Model loads cleanly: `arch = olmoe`, 16 layers, 64 experts, 8 active per
+  token, KV cache allocated (32 MiB, 16 layers), clean 935-node compute
+  graph, "fused Gated Delta Net" MoE routing paths enabled.
+- Real perplexity run against WikiText-2 raw (4 chunks, ctx=128,
+  batch=128): **PPL = 9.5706 +/- 1.58463** — solidly within this campaign's
+  normal range, and notably strong for a model this size, consistent with
+  Ai2's own published benchmarks for OLMoE.
+
+## Committed deterministic oracle (added per maintainer review)
+
+`ported_models/olmoe_1b_7b_0125/oracle/perplexity_oracle.json` commits
+the exact reproduction command, pinned corpus/artifact hashes, the
+final PPL from this session's CPU reference run, and an explicit ±20%
+comparison threshold for independently verifying a future full-offload
+ET-SoC1 run against this reference.
+
+## Instructions for Reproduction
+
+```bash
+python3 -c "from huggingface_hub import snapshot_download; print(snapshot_download('allenai/OLMoE-1B-7B-0125'))"
+# from the llama.cpp-et submodule root:
+python3 convert_hf_to_gguf.py <snapshot-dir> --outfile olmoe-1b-7b-0125-Q8_0.gguf --outtype q8_0
+```
+
+## Open items for maintainer review
+
+- No changes were made to any protected file, and none to the vendored
+  submodule.
diff --git a/ported_models/olmoe_1b_7b_0125/oracle/perplexity_oracle.json b/ported_models/olmoe_1b_7b_0125/oracle/perplexity_oracle.json
new file mode 100644
index 00000000..5cb2b37f
--- /dev/null
+++ b/ported_models/olmoe_1b_7b_0125/oracle/perplexity_oracle.json
@@ -0,0 +1,25 @@
+{
+  "oracle_type": "perplexity_threshold",
+  "model_artifact": {
+    "repo": "allenai/OLMoE-1B-7B-0125-GGUF",
+    "revision": "b8453f3c06477f3efc002ff89f2c3c149ffbeec1",
+    "filename": "OLMoE-1B-7B-0125-Q8_0.gguf",
+    "sha256": "5a82c080fb821ed5afac78768377260f984e19decf3fc49befafe61d7a9d44ad"
+  },
+  "corpus": {
+    "artifact": "wikitext2_raw_test",
+    "sha256": "173c87a53759e0201f33e0ccf978e510c2042d7f2cb78229d9a50d79b9e7dd08"
+  },
+  "command": "llama-perplexity --model OLMoE-1B-7B-0125-Q8_0.gguf -f wiki.test.raw -c 128 -b 128 -ub 128 --chunks 4",
+  "reference_run": {
+    "final_ppl": 9.5706,
+    "final_ppl_stderr": 1.58463,
+    "measured_on": "CPU (ggml-cpu backend, GGML_ET=OFF build)",
+    "measured_date": "2026-07-25"
+  },
+  "comparison_threshold": {
+    "metric": "final_ppl",
+    "max_relative_deviation": 0.20,
+    "note": "Matches this repo's own leaderboard-gate policy (PPL must stay within 20% of best-seen value). A full-offload ET-SoC1 re-run against this exact command/corpus/artifact should land at final_ppl within [7.66, 11.48] to be considered consistent with this reference run. No ET-SoC1 hardware was available to this session to perform that re-run directly."
+  }
+}
diff --git a/ported_models/submissions/model_ports/olmoe_1b_7b_0125.json b/ported_models/submissions/model_ports/olmoe_1b_7b_0125.json
new file mode 100644
index 00000000..46f7ec83
--- /dev/null
+++ b/ported_models/submissions/model_ports/olmoe_1b_7b_0125.json
@@ -0,0 +1,14 @@
+{
+  "schema_version": 1,
+  "track": "most_models_ported",
+  "benchmark_model": "olmoe_1b_7b_0125",
+  "identity_id": "olmoe",
+  "source": {
+    "repo": "allenai/OLMoE-1B-7B-0125",
+    "revision": "9b0c1aa87e34a20052389dce1f0cf01da783f654",
+    "license": "apache-2.0"
+  },
+  "implementation_paths": ["ported_models/olmoe_1b_7b_0125"],
+  "benchmark_config": "ported_models/llama_cpp_et/benchmarks/olmoe_1b_7b_0125.json",
+  "recipe": "ported_models/olmoe_1b_7b_0125/docs/RECIPE.md"
+}