{ "schemaVersion": 1, "operationId": "20260828085933-15afea347d", "checkedAt": "2026-08-28T17:10:00+08:00", "status": "pinned-source-zero-model-call-glm52-vllm-offline-batch-audit", "revisions": { "vllm": "0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665", "currentVllmDocs": "5f213ed1592903b7bc38f173d320dac1b2769303", "recipe": "5943215a27acb4a243e9d27bd69daf491034cfea", "glm52Base": "b4734de4facf877f85769a911abafc5283eab3d9", "glm52Fp8": "ba978f7d347eaf65d22f1a86833408afdb953541" }, "sourceReceipts": [ { "id": "vllm-run-batch-source", "url": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/vllm/entrypoints/openai/run_batch.py", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/vllm/entrypoints/openai/run_batch.py", "httpStatus": 200, "bytes": 31032, "sha256": "49d78652a4ff6ed797264154cfe47c25aa8f259f3ca08be9f780b6a2b9aa1dff" }, { "id": "vllm-openai-batch-readme", "url": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/examples/features/openai_batch/README.md", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/examples/features/openai_batch/README.md", "httpStatus": 200, "bytes": 13387, "sha256": "60386f81357fbd04c17004b07d5d2431ab9ae282b7cc14e4881a250cabf0b9ff" }, { "id": "vllm-openai-batch-example", "url": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/examples/features/openai_batch/openai_example_batch.jsonl", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/examples/features/openai_batch/openai_example_batch.jsonl", "httpStatus": 200, "bytes": 573, "sha256": "66fdb813bb35544f6fc18042c692dfa1b863e04066ffe9910e7a64139dff1006" }, { "id": "vllm-engine-arguments", "url": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/vllm/engine/arg_utils.py", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/vllm/engine/arg_utils.py", "httpStatus": 200, "bytes": 111675, "sha256": "3ade2488897d08b5c1fd4ddb815eb5743169907d59741a47867185caab7f250d" }, { "id": "vllm-frontend-arguments", "url": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/vllm/entrypoints/openai/cli_args.py", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/vllm/entrypoints/openai/cli_args.py", "httpStatus": 200, "bytes": 18047, "sha256": "b10da703f00977dcbbd379f6e0898087a8036693258c4d0c066a2a99b19b0334" }, { "id": "vllm-v023-supported-models", "url": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/docs/models/supported_models.md", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/docs/models/supported_models.md", "httpStatus": 200, "bytes": 59570, "sha256": "43eca89f124731e844d09d3af80fd0769969c3beda7483abdb2549d0210c18a4" }, { "id": "vllm-current-supported-models", "url": "https://raw.githubusercontent.com/vllm-project/vllm/5f213ed1592903b7bc38f173d320dac1b2769303/docs/models/supported_models.md", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/5f213ed1592903b7bc38f173d320dac1b2769303/docs/models/supported_models.md", "httpStatus": 200, "bytes": 59436, "sha256": "2c4de9536f7e7cea071e971690331186fc95a19dd07c5fb48b5bd8d693a7fac5" }, { "id": "vllm-batch-invariance-example", "url": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/examples/features/batch_invariance/reproducibility_offline.py", "finalUrl": "https://raw.githubusercontent.com/vllm-project/vllm/0fc695fc6d1d82e9a5ac6835ac8e4e1c83703665/examples/features/batch_invariance/reproducibility_offline.py", "httpStatus": 200, "bytes": 1339, "sha256": "49612f00806c0de8e4ff0044cf9bd06f1e3580a15485126664dee791613358c8" }, { "id": "vllm-glm52-recipe", "url": "https://raw.githubusercontent.com/vllm-project/recipes/5943215a27acb4a243e9d27bd69daf491034cfea/models/zai-org/GLM-5.2.yaml", "finalUrl": "https://raw.githubusercontent.com/vllm-project/recipes/5943215a27acb4a243e9d27bd69daf491034cfea/models/zai-org/GLM-5.2.yaml", "httpStatus": 200, "bytes": 15247, "sha256": "517662028453e8a0b62f53a8b507d6b8ac502aaecbd5452b396c37ea2fbae4a7" }, { "id": "hf-base-revision-api", "url": "https://huggingface.co/api/models/zai-org/GLM-5.2/revision/b4734de4facf877f85769a911abafc5283eab3d9", "finalUrl": "https://huggingface.co/api/models/zai-org/GLM-5.2/revision/b4734de4facf877f85769a911abafc5283eab3d9", "httpStatus": 200, "bytes": 23596, "sha256": "e54d8c5f3934f2ed7b33871ca36b5912a95500d72d427334d44536debc03069f" }, { "id": "hf-base-config", "url": "https://huggingface.co/zai-org/GLM-5.2/resolve/b4734de4facf877f85769a911abafc5283eab3d9/config.json", "finalUrl": "https://huggingface.co/api/resolve-cache/models/zai-org/GLM-5.2/b4734de4facf877f85769a911abafc5283eab3d9/config.json?%2Fzai-org%2FGLM-5.2%2Fresolve%2Fb4734de4facf877f85769a911abafc5283eab3d9%2Fconfig.json=&etag=%22f8b342fa9e8fecdc667312641760b4a3015b5e4f%22", "httpStatus": 200, "bytes": 3732, "sha256": "185f93ee6d12548e16a847e279dc0c3c90b1524c970b0866b42fb545747d859a" }, { "id": "hf-generation-config", "url": "https://huggingface.co/zai-org/GLM-5.2/resolve/b4734de4facf877f85769a911abafc5283eab3d9/generation_config.json", "finalUrl": "https://huggingface.co/api/resolve-cache/models/zai-org/GLM-5.2/b4734de4facf877f85769a911abafc5283eab3d9/generation_config.json?%2Fzai-org%2FGLM-5.2%2Fresolve%2Fb4734de4facf877f85769a911abafc5283eab3d9%2Fgeneration_config.json=&etag=%22216bbb092d0df0dda1dc4cedd789b92286c4c001%22", "httpStatus": 200, "bytes": 194, "sha256": "ac76b43d8683d3b930126870fc8be73d8679308fe752fa1f381096d8354f6a55" }, { "id": "hf-chat-template", "url": "https://huggingface.co/zai-org/GLM-5.2/resolve/b4734de4facf877f85769a911abafc5283eab3d9/chat_template.jinja", "finalUrl": "https://huggingface.co/api/resolve-cache/models/zai-org/GLM-5.2/b4734de4facf877f85769a911abafc5283eab3d9/chat_template.jinja?%2Fzai-org%2FGLM-5.2%2Fresolve%2Fb4734de4facf877f85769a911abafc5283eab3d9%2Fchat_template.jinja=&etag=%22672034e3a1196d8b9fc0116ddc037cb554227f89%22", "httpStatus": 200, "bytes": 5076, "sha256": "172dc74a35e1752df75ecfb2b2cf9326d2852bb1379868ebeec9571654489679" }, { "id": "hf-fp8-revision-api", "url": "https://huggingface.co/api/models/zai-org/GLM-5.2-FP8/revision/ba978f7d347eaf65d22f1a86833408afdb953541", "finalUrl": "https://huggingface.co/api/models/zai-org/GLM-5.2-FP8/revision/ba978f7d347eaf65d22f1a86833408afdb953541", "httpStatus": 200, "bytes": 35738, "sha256": "03597386bf1e39b375c2445991ab6da1fe7da4a5643b27a9ce16731fe2ec4dc2" }, { "id": "hf-fp8-config", "url": "https://huggingface.co/zai-org/GLM-5.2-FP8/resolve/ba978f7d347eaf65d22f1a86833408afdb953541/config.json", "finalUrl": "https://huggingface.co/api/resolve-cache/models/zai-org/GLM-5.2-FP8/ba978f7d347eaf65d22f1a86833408afdb953541/config.json?%2Fzai-org%2FGLM-5.2-FP8%2Fresolve%2Fba978f7d347eaf65d22f1a86833408afdb953541%2Fconfig.json=&etag=%224e1f0168afd127189fb1c4ddb1d4476a4fca96ac%22", "httpStatus": 200, "bytes": 29464, "sha256": "22e49334abf8562fecf70ca3292ba3f5b33f5602fb2bf10b52dd64a66cfe65ff" }, { "id": "zai-glm52-release", "url": "https://z.ai/blog/glm-5.2", "finalUrl": "https://z.ai/blog/glm-5.2", "httpStatus": 200, "bytes": 598, "sha256": "a9e8c2b6f34717d69e3a0aa26bb117256a4d8c95bd299910c2a693000ee88fe8" }, { "id": "zhipu-research-index", "url": "https://www.zhipuai.cn/zh/research", "finalUrl": "https://www.zhipuai.cn/zh/research", "httpStatus": 200, "bytes": 1235119, "sha256": "7d50f4c290fbc240f50fabc8d18f7a499688f968f53951655572909ddf547d4e" } ], "sourceContract": { "openAiBatchIsJsonl": true, "customIdIsDeveloperJoinKey": true, "postMethodDocumented": true, "chatCompletionsValidated": true, "streamRejected": true, "concurrentSubmission": true, "localInputAndOutputFlags": true, "modelRevisionArgumentAvailable": true, "reasoningParserArgumentAvailable": true, "frontendThinkingKwargsAvailable": true, "v023ModelTableOmittedGlm52": true, "currentModelTableListsGlm52": true, "batchInvariantFlagDocumented": true }, "modelContract": { "baseRevision": "b4734de4facf877f85769a911abafc5283eab3d9", "fp8Revision": "ba978f7d347eaf65d22f1a86833408afdb953541", "basePublic": true, "fp8Public": true, "license": "mit", "architecture": "GlmMoeDsaForCausalLM", "modelType": "glm_moe_dsa", "maxPositionEmbeddings": 1048576, "totalLayers": 78, "quantizationMethod": "fp8", "quantizationFormat": "e4m3", "eosTokenIds": [ 154820, 154827, 154829 ], "defaultTemperature": 1, "defaultTopP": 0.95, "templateDefaultEffortIsMax": true, "templateSupportsNonThinking": true, "recipeMinimumVllm": "0.23.0", "recipeDefaultModel": "zai-org/GLM-5.2-FP8", "recipeFp8MinimumVramGb": 893 }, "requestFixture": { "requestCount": 4, "customIds": [ "extract-0001", "review-0002", "classify-0003", "plan-0004" ], "nonThinkingCount": 2, "reasoningCount": 2, "plannedCompletionTokens": 1472, "jsonlBytes": 1630, "sha256": "85fda02184b27644e763f689299061c93ae98fdaf0fd48ab6ed647b0b63b865d", "validationErrors": [] }, "invalidCaseResults": [ { "label": "duplicate-custom-id", "mutation": "duplicate-first-request", "expectedError": "duplicate-custom-id", "actualErrors": [ "duplicate-custom-id" ], "rejectedAsExpected": true }, { "label": "wrong-method", "mutation": "method-get", "expectedError": "method-must-be-post", "actualErrors": [ "method-must-be-post" ], "rejectedAsExpected": true }, { "label": "wrong-endpoint", "mutation": "url-responses", "expectedError": "unsupported-endpoint", "actualErrors": [ "unsupported-endpoint" ], "rejectedAsExpected": true }, { "label": "mixed-model", "mutation": "model-base-checkpoint", "expectedError": "model-mismatch", "actualErrors": [ "model-mismatch" ], "rejectedAsExpected": true }, { "label": "streaming-request", "mutation": "stream-true", "expectedError": "streaming-not-allowed", "actualErrors": [ "streaming-not-allowed" ], "rejectedAsExpected": true }, { "label": "implicit-thinking-default", "mutation": "remove-thinking-policy", "expectedError": "thinking-policy-must-be-explicit", "actualErrors": [ "thinking-policy-must-be-explicit" ], "rejectedAsExpected": true }, { "label": "zero-output-budget", "mutation": "max-completion-zero", "expectedError": "invalid-max-completion-tokens", "actualErrors": [ "invalid-max-completion-tokens" ], "rejectedAsExpected": true } ], "syntheticResultFixture": { "jsonlBytes": 1601, "sha256": "67aeabac56978e39c43b5adeb5a7d36a42d2fcd0c8675418cc99a0265af362e6", "syntheticOnly": true, "modelGeneratedTokens": 0 }, "outputAudit": { "inputRequestCount": 4, "outputResultCount": 4, "duplicateIds": [], "unknownIds": [], "missingIds": [], "resultsWereShuffled": true, "joinedByCustomId": true, "acceptedIds": [ "extract-0001", "plan-0004" ], "quarantinedIds": [ "classify-0003", "review-0002" ], "usage": { "promptTokens": 568, "completionTokens": 704, "totalTokens": 1272 }, "rows": [ { "customId": "plan-0004", "statusCode": 200, "finishReason": "stop", "promptTokens": 200, "completionTokens": 180, "disposition": "accept", "reasons": [] }, { "customId": "classify-0003", "statusCode": 400, "finishReason": null, "promptTokens": 0, "completionTokens": 0, "disposition": "quarantine", "reasons": [ "non-200-status", "non-null-error", "missing-response-body", "missing-choice" ] }, { "customId": "extract-0001", "statusCode": 200, "finishReason": "stop", "promptTokens": 48, "completionTokens": 12, "disposition": "accept", "reasons": [] }, { "customId": "review-0002", "statusCode": 200, "finishReason": "length", "promptTokens": 320, "completionTokens": 512, "disposition": "quarantine", "reasons": [ "finish-reason-length" ] } ] }, "overlapAudit": { "prepublicationCanonicalCount": 94, "registeredContentPages": 87, "distinctIntent": true, "nearestPages": [ { "canonical": "https://glm52.ai/guides/run-glm-5-2-locally/", "ownedJob": "choose hardware, checkpoint, and local serving route", "newBoundary": "does not own OpenAI-style JSONL batch validation or result reconciliation" }, { "canonical": "https://glm52.ai/guides/glm-5-2-vllm-thinking-token-budget/", "ownedJob": "reserve visible answer capacity for one self-hosted vLLM request", "newBoundary": "does not own multi-request files, custom ID joins, or batch release gates" }, { "canonical": "https://glm52.ai/guides/glm-5-2-prompt-caching/", "ownedJob": "structure repeated prefixes and calculate hosted API cache cost", "newBoundary": "does not own offline runner input and output receipts" } ] }, "searchSupply": { "query": "GLM-5.2 vLLM offline batch inference JSONL", "serpApiRequests": 1, "requestId": "serpapi-864ccd5a53dc4d74aa77f50bec6ef68b", "requestCost": 1, "knownMonthlyUsageAfter": 819, "status": "failed", "evaluation": "failed", "retried": false, "decisionChanged": false }, "runtimeBoundary": { "modelCalls": 0, "localGpuRuns": 0, "modelWeightBytesDownloaded": 0, "vllmProcessesStarted": 0, "thirdPartyCliRuns": 0, "fixtureRuntime": "Node.js static source and JSONL contract audit only" }, "aiHot": [ { "id": "cmtcpvj9e04cpro645lmugnew", "permalink": "https://aihot.virxact.com/items/cmtcpvj9e04cpro645lmugnew", "classification": "weak-glm-link", "note": "Generic Groq and Colab notebooks do not establish a GLM-5.2 offline batch contract." }, { "id": "cmtcjzlxy03f8rodbxqdotbhg", "permalink": "https://aihot.virxact.com/items/cmtcjzlxy03f8rodbxqdotbhg", "classification": "weak-glm-link", "note": "The Hy4 preview concerns another model family and does not answer a GLM-5.2 operator job." }, { "id": "cmtc61ej101lqrojqto587wu9", "permalink": "https://aihot.virxact.com/items/cmtc61ej101lqrojqto587wu9", "classification": "weak-glm-link", "note": "A Midjourney image editor has no direct GLM-5.2 text-batch workflow." }, { "id": "cmtc5i81401mlroosz6bsplf6", "permalink": "https://aihot.virxact.com/items/cmtc5i81401mlroosz6bsplf6", "classification": "weak-glm-link", "note": "Gemini transcription is a different model, modality, and API task." }, { "id": "cmtc4fn8n01mwrozaq4bot3b4", "permalink": "https://aihot.virxact.com/items/cmtc4fn8n01mwrozaq4bot3b4", "classification": "weak-glm-link", "note": "A revenue forecast supplies no model-specific inference decision." }, { "id": "cmtc05bnj015srome8fm42xy8", "permalink": "https://aihot.virxact.com/items/cmtc05bnj015srome8fm42xy8", "classification": "weak-glm-link", "note": "The xAI litigation report is unrelated to a safe, verifiable GLM-5.2 reader task." }, { "id": "cmtc7nham01clrodmgnklmfp8", "permalink": "https://aihot.virxact.com/items/cmtc7nham01clrodmgnklmfp8", "classification": "weak-glm-link", "note": "Claude account-key administration is unrelated to GLM-5.2 offline execution." } ], "decision": { "workflow": "preflight-jsonl-then-run-pinned-vllm-batch-then-reconcile-by-custom-id-and-quarantine-any-non-stop-or-error-row", "inputRule": "One JSON object per line, one unique custom_id per request, one pinned model, POST chat/completions only, streaming off, and an explicit thinking policy.", "outputRule": "Release only rows with the expected custom_id, HTTP 200, null error, a non-empty choice, finish_reason=stop, and internally consistent usage totals.", "truncationRule": "finish_reason=length is a failed job for release purposes even when content is non-empty.", "reproducibilityRule": "Hash the canonical input and output files, pin revisions and engine settings, and canary batch invariance; temperature zero alone is not a bitwise reproducibility guarantee." }, "invariants": { "seventeenPrimarySourceReceipts": true, "allSourcesReturned200": true, "allSourceHashesValid": true, "allPinnedSourceMarkersPresent": true, "immutableModelRevisionsMatch": true, "glm52Fp8ContractMatches": true, "recipeBoundaryMatches": true, "validFixturePasses": true, "sevenInvalidCasesRejected": true, "fourUniqueRequests": true, "explicitThinkingSplit": true, "plannedTokenCeilingMatches": true, "outputJoinIsComplete": true, "shuffledOrderWasHandled": true, "twoAcceptedTwoQuarantined": true, "lengthRowWasQuarantined": true, "usageArithmeticMatches": true, "distinctIntent": true, "oneFailedSerpRequestNoRetry": true, "allAiHotClassifiedWeak": true, "zeroModelGpuAndWeightWork": true } }