diff --git a/data/artificial-analysis-index.json b/data/artificial-analysis-index.json index 1c734009..169925a1 100644 --- a/data/artificial-analysis-index.json +++ b/data/artificial-analysis-index.json @@ -4,7 +4,7 @@ "sourceUrl": "https://artificialanalysis.ai/leaderboards/models", "methodologyUrl": "https://artificialanalysis.ai/methodology/intelligence-benchmarking", "indexVersion": "4.1.1", - "observedAt": "2026-08-18", + "observedAt": "2026-08-19", "legacyMissingModelIds": ["cursor-composer-2", "cursor-composer-2-5"], "entries": [ { @@ -355,6 +355,12 @@ "estimated": false, "configuration": "GLM-5.2 (max)" }, + { + "modelId": "glm-5-3", + "score": 60, + "estimated": false, + "configuration": "GLM-5.3 (max)" + }, { "modelId": "gpt-4-1", "score": 20, diff --git a/data/data-health.json b/data/data-health.json index c4da8918..b677cdb0 100644 --- a/data/data-health.json +++ b/data/data-health.json @@ -1,5 +1,5 @@ { - "asOf": "2026-08-18", + "asOf": "2026-08-19", "thresholds": { "models": 30, "providers": 30, @@ -10,14 +10,14 @@ "vendors": 90 }, "summary": { - "totalRecords": 261, - "recordsWithSources": 261, - "verifiedRecords": 261, - "provenanceComplete": 261, + "totalRecords": 262, + "recordsWithSources": 262, + "verifiedRecords": 262, + "provenanceComplete": 262, "staleVerifiedRecords": 2, "translationPlaceholderValues": 300, "danglingRelationships": 0, - "modelBenchmarkCoverage": 9, + "modelBenchmarkCoverage": 8.9, "productsWithPricing": 66, "productRecords": 67, "communityUrlsPopulated": 331, @@ -53,9 +53,9 @@ "stale": 0 }, "models": { - "total": 130, - "verified": 130, - "provenanceComplete": 130, + "total": 131, + "verified": 131, + "provenanceComplete": 131, "stale": 2 }, "providers": { @@ -134,14 +134,14 @@ "code": "stale-verification", "category": "models", "id": "claude-haiku-4-5", - "message": "Last reviewed 31 days ago; threshold is 30 days." + "message": "Last reviewed 32 days ago; threshold is 30 days." }, { "severity": "warning", "code": "stale-verification", "category": "models", "id": "gpt-5-2", - "message": "Last reviewed 31 days ago; threshold is 30 days." + "message": "Last reviewed 32 days ago; threshold is 30 days." } ] } diff --git a/data/model-price-intelligence-index.json b/data/model-price-intelligence-index.json index 19f0b2ab..b0264e76 100644 --- a/data/model-price-intelligence-index.json +++ b/data/model-price-intelligence-index.json @@ -58,7 +58,7 @@ "linearLabelAnchor": "start" }, { - "modelId": "glm-5-2", + "modelId": "glm-5-3", "labelDx": -22, "labelDy": -18, "labelAnchor": "end" diff --git a/docs/DATA-HEALTH.md b/docs/DATA-HEALTH.md index fb1d3ab2..fa43e5e3 100644 --- a/docs/DATA-HEALTH.md +++ b/docs/DATA-HEALTH.md @@ -1,19 +1,19 @@ # Data Health Report -Snapshot date: 2026-08-18. Regenerate with `pnpm data-health:report`. +Snapshot date: 2026-08-19. Regenerate with `pnpm data-health:report`. ## Scorecard | Metric | Value | | --- | ---: | -| Manifest records | 261 | -| Records with structured sources | 261 | -| Verified records | 261 | -| Verified with complete provenance | 261 | +| Manifest records | 262 | +| Records with structured sources | 262 | +| Verified records | 262 | +| Verified with complete provenance | 262 | | Stale verified records | 2 | | Non-English values identical to English | 300 | | Dangling product relationships | 0 | -| Model benchmark coverage | 9% | +| Model benchmark coverage | 8.9% | | Products with pricing | 66/67 | | Community URLs with provenance | 331/331 | | Duplicated vendor community URLs | 0 | @@ -27,7 +27,7 @@ Snapshot date: 2026-08-18. Regenerate with `pnpm data-health:report`. | clis | 28 | 28 | 28 | 0 | | desktops | 12 | 12 | 12 | 0 | | extensions | 18 | 18 | 18 | 0 | -| models | 130 | 130 | 130 | 2 | +| models | 131 | 131 | 131 | 2 | | providers | 17 | 17 | 17 | 0 | | vendors | 47 | 47 | 47 | 0 | @@ -62,8 +62,8 @@ visible in the scorecards and `data/data-health.json`. | Severity | Issue | Record | Detail | | --- | --- | --- | --- | -| warning | stale-verification | models/claude-haiku-4-5 | Last reviewed 31 days ago; threshold is 30 days. | -| warning | stale-verification | models/gpt-5-2 | Last reviewed 31 days ago; threshold is 30 days. | +| warning | stale-verification | models/claude-haiku-4-5 | Last reviewed 32 days ago; threshold is 30 days. | +| warning | stale-verification | models/gpt-5-2 | Last reviewed 32 days ago; threshold is 30 days. | ## Freshness Thresholds diff --git a/manifests/models/glm-5-2.json b/manifests/models/glm-5-2.json index 154c75f4..52c73781 100644 --- a/manifests/models/glm-5-2.json +++ b/manifests/models/glm-5-2.json @@ -2,40 +2,40 @@ "$schema": "../$schemas/model.schema.json", "id": "glm-5-2", "name": "GLM-5.2", - "description": "Z.ai's open-weight flagship for long-horizon agentic coding, flexible reasoning effort, and stable work across a 1M-token context window.", + "description": "Z.ai's open-weight model for long-horizon agentic coding, flexible reasoning effort, and stable work across a 1M-token context window.", "translations": { "de": { - "description": "Z.ais offenes Flaggschiff für langfristige agentische Programmierung, flexible Denkleistung und stabile Arbeit über ein Kontextfenster von 1 Mio. Token." + "description": "Z.ais offenes Modell für langfristige agentische Programmierung, flexible Denkleistung und stabile Arbeit über ein Kontextfenster von 1 Mio. Token." }, "es": { - "description": "El modelo insignia de pesos abiertos de Z.ai para programación agéntica prolongada, razonamiento flexible y trabajo estable con contexto de 1 millón de tokens." + "description": "El modelo de pesos abiertos de Z.ai para programación agéntica prolongada, razonamiento flexible y trabajo estable con contexto de 1 millón de tokens." }, "fr": { - "description": "Le modèle phare à poids ouverts de Z.ai pour le codage agentique long, l’effort de raisonnement flexible et un travail stable sur 1 million de jetons." + "description": "Le modèle à poids ouverts de Z.ai pour le codage agentique long, l’effort de raisonnement flexible et un travail stable sur 1 million de jetons." }, "id": { - "description": "Model unggulan open-weight Z.ai untuk coding agentik jangka panjang, upaya penalaran fleksibel, dan kerja stabil dalam konteks 1 juta token." + "description": "Model open-weight Z.ai untuk coding agentik jangka panjang, upaya penalaran fleksibel, dan kerja stabil dalam konteks 1 juta token." }, "ja": { - "description": "長期のエージェント型コーディング、柔軟な推論強度、100万トークンのコンテキストにわたる安定した作業に対応するZ.aiのオープンウェイト旗艦モデル。" + "description": "長期のエージェント型コーディング、柔軟な推論強度、100万トークンのコンテキストにわたる安定した作業に対応するZ.aiのオープンウェイトモデル。" }, "ko": { - "description": "장기 에이전트 코딩, 유연한 추론 강도 및 100만 토큰 컨텍스트 전반의 안정적인 작업을 위한 Z.ai의 오픈 웨이트 플래그십 모델입니다." + "description": "장기 에이전트 코딩, 유연한 추론 강도 및 100만 토큰 컨텍스트 전반의 안정적인 작업을 위한 Z.ai의 오픈 웨이트 모델입니다." }, "pt": { - "description": "O modelo principal de pesos abertos da Z.ai para programação agêntica prolongada, raciocínio flexível e trabalho estável em contexto de 1 milhão de tokens." + "description": "O modelo de pesos abertos da Z.ai para programação agêntica prolongada, raciocínio flexível e trabalho estável em contexto de 1 milhão de tokens." }, "ru": { - "description": "Открытая флагманская модель Z.ai для длительного агентного программирования, гибкой глубины рассуждений и устойчивой работы с контекстом 1 млн токенов." + "description": "Открытая модель Z.ai для длительного агентного программирования, гибкой глубины рассуждений и устойчивой работы с контекстом 1 млн токенов." }, "tr": { - "description": "Z.ai'ın uzun süreli ajan tabanlı kodlama, esnek akıl yürütme düzeyi ve 1 milyon token bağlamda kararlı çalışma için açık ağırlıklı amiral modeli." + "description": "Z.ai'ın uzun süreli ajan tabanlı kodlama, esnek akıl yürütme düzeyi ve 1 milyon token bağlamda kararlı çalışma için açık ağırlıklı modeli." }, "zh-Hans": { - "description": "Z.ai 面向长周期智能体编码、灵活推理强度和 100 万 token 上下文稳定工作的开放权重旗舰模型。" + "description": "Z.ai 面向长周期智能体编码、灵活推理强度和 100 万 token 上下文稳定工作的开放权重模型。" }, "zh-Hant": { - "description": "Z.ai 面向長週期智慧體程式設計、彈性推理強度和 100 萬 token 上下文穩定工作的開放權重旗艦模型。" + "description": "Z.ai 面向長週期智慧體程式設計、彈性推理強度和 100 萬 token 上下文穩定工作的開放權重模型。" } }, "verified": true, @@ -80,7 +80,7 @@ "fields": ["size", "platformUrls"] } ], - "lastVerifiedAt": "2026-07-21", + "lastVerifiedAt": "2026-08-19", "verifiedBy": "codex-agent", "confidence": "high", "websiteUrl": "https://z.ai", @@ -116,7 +116,7 @@ ] }, "releaseDate": "2026-06-16", - "lifecycle": "latest", + "lifecycle": "maintained", "knowledgeCutoff": null, "inputModalities": ["text"], "outputModalities": ["text"], diff --git a/manifests/models/glm-5-3.json b/manifests/models/glm-5-3.json new file mode 100644 index 00000000..b6eeb87d --- /dev/null +++ b/manifests/models/glm-5-3.json @@ -0,0 +1,141 @@ +{ + "$schema": "../$schemas/model.schema.json", + "id": "glm-5-3", + "name": "GLM-5.3", + "description": "Z.ai's latest flagship model for complex software engineering and long-horizon agent tasks, with always-on reasoning and a 1M-token context window.", + "translations": { + "de": { + "description": "Z.ais neuestes Flaggschiffmodell für komplexe Softwareentwicklung und langfristige Agentenaufgaben, mit stets aktivem Reasoning und einem Kontextfenster von 1 Mio. Token." + }, + "es": { + "description": "El modelo insignia más reciente de Z.ai para ingeniería de software compleja y tareas agénticas prolongadas, con razonamiento siempre activo y un contexto de 1 millón de tokens." + }, + "fr": { + "description": "Le dernier modèle phare de Z.ai pour l’ingénierie logicielle complexe et les tâches agentiques longues, avec raisonnement permanent et contexte de 1 million de jetons." + }, + "id": { + "description": "Model unggulan terbaru Z.ai untuk rekayasa perangkat lunak kompleks dan tugas agentik jangka panjang, dengan penalaran selalu aktif dan konteks 1 juta token." + }, + "ja": { + "description": "複雑なソフトウェア開発と長期エージェントタスク向けのZ.ai最新旗艦モデル。常時有効な推論と100万トークンのコンテキストに対応します。" + }, + "ko": { + "description": "복잡한 소프트웨어 엔지니어링과 장기 에이전트 작업을 위한 Z.ai의 최신 플래그십 모델로, 상시 추론과 100만 토큰 컨텍스트를 지원합니다." + }, + "pt": { + "description": "O modelo principal mais recente da Z.ai para engenharia de software complexa e tarefas agênticas prolongadas, com raciocínio sempre ativo e contexto de 1 milhão de tokens." + }, + "ru": { + "description": "Новейшая флагманская модель Z.ai для сложной программной инженерии и длительных агентных задач с постоянно включённым рассуждением и контекстом в 1 млн токенов." + }, + "tr": { + "description": "Z.ai'ın karmaşık yazılım mühendisliği ve uzun süreli ajan görevleri için en yeni amiral modeli; sürekli akıl yürütme ve 1 milyon token bağlam sunar." + }, + "zh-Hans": { + "description": "Z.ai 面向复杂软件工程和长周期智能体任务的最新旗舰模型，支持始终开启的推理和 100 万 token 上下文。" + }, + "zh-Hant": { + "description": "Z.ai 面向複雜軟體工程和長週期智慧體任務的最新旗艦模型，支援始終開啟的推理和 100 萬 token 上下文。" + } + }, + "verified": true, + "sources": [ + { + "url": "https://docs.z.ai/guides/llm/glm-5.3", + "title": "GLM-5.3 model documentation", + "fields": [ + "name", + "description", + "websiteUrl", + "docsUrl", + "size", + "activeParameters", + "contextWindow", + "maxOutput", + "inputModalities", + "outputModalities", + "capabilities" + ] + }, + { + "url": "https://huggingface.co/zai-org/GLM-5.2", + "title": "Official GLM-5.2 base checkpoint and model card", + "fields": ["size"] + }, + { + "url": "https://docs.z.ai/guides/llm/glm-5", + "title": "Official GLM-5 architecture parameter disclosure", + "fields": ["activeParameters"] + }, + { + "url": "https://docs.z.ai/release-notes/new-released", + "title": "Z.ai model release notes", + "fields": ["releaseDate", "lifecycle"] + }, + { + "url": "https://docs.z.ai/guides/overview/pricing", + "title": "Z.ai API pricing", + "fields": ["tokenPricing"] + }, + { + "url": "https://artificialanalysis.ai/models/glm-5-3", + "title": "Artificial Analysis GLM-5.3 model page", + "fields": ["platformUrls.artificialAnalysis"] + } + ], + "lastVerifiedAt": "2026-08-19", + "verifiedBy": "codex-agent", + "confidence": "high", + "websiteUrl": "https://z.ai", + "docsUrl": "https://docs.z.ai/guides/llm/glm-5.3", + "vendor": "Z.ai", + "size": "753B", + "activeParameters": "40B", + "contextWindow": 1000000, + "maxOutput": 131072, + "tokenPricing": { + "status": "available", + "primaryOffer": "global-standard", + "offers": [ + { + "id": "global-standard", + "currency": "USD", + "region": "global", + "serviceTier": "standard", + "effectiveFrom": null, + "effectiveTo": null, + "tiers": [ + { + "condition": null, + "rates": { + "input": 1.4, + "output": 4.4, + "cacheRead": 0.26, + "cacheWrite": null + } + } + ] + } + ] + }, + "releaseDate": "2026-08-18", + "lifecycle": "latest", + "knowledgeCutoff": null, + "inputModalities": ["text"], + "outputModalities": ["text"], + "capabilities": ["function-calling", "structured-outputs", "reasoning"], + "benchmarks": { + "sweBench": null, + "terminalBench": null, + "mmmu": null, + "mmmuPro": null, + "webDevArena": null, + "sciCode": null, + "liveCodeBench": null + }, + "platformUrls": { + "huggingface": null, + "artificialAnalysis": "https://artificialanalysis.ai/models/glm-5-3", + "openrouter": null + } +} diff --git a/manifests/vendors/z-ai.json b/manifests/vendors/z-ai.json index 3529d81a..2f114c5b 100644 --- a/manifests/vendors/z-ai.json +++ b/manifests/vendors/z-ai.json @@ -11,7 +11,7 @@ { "id": "glm", "name": "GLM", - "modelIds": ["glm-4-5", "glm-4-6", "glm-4-7", "glm-5", "glm-5-1", "glm-5-2"] + "modelIds": ["glm-4-5", "glm-4-6", "glm-4-7", "glm-5", "glm-5-1", "glm-5-2", "glm-5-3"] }, { "id": "glm-air-flash", diff --git a/src/lib/generated/metadata.ts b/src/lib/generated/metadata.ts index 38ca7a5a..d2745786 100644 --- a/src/lib/generated/metadata.ts +++ b/src/lib/generated/metadata.ts @@ -3639,7 +3639,7 @@ export const stackCounts: Record = { clis: 28, desktops: 12, extensions: 18, - models: 130, + models: 131, 'model-providers': 17, vendors: 47, } diff --git a/src/lib/generated/models.ts b/src/lib/generated/models.ts index 6a76434e..42761fcb 100644 --- a/src/lib/generated/models.ts +++ b/src/lib/generated/models.ts @@ -62,6 +62,7 @@ import Glm47Flash from '../../../manifests/models/glm-4-7-flash.json' import Glm5 from '../../../manifests/models/glm-5.json' import Glm51 from '../../../manifests/models/glm-5-1.json' import Glm52 from '../../../manifests/models/glm-5-2.json' +import Glm53 from '../../../manifests/models/glm-5-3.json' import Glm5Turbo from '../../../manifests/models/glm-5-turbo.json' import Glm5vTurbo from '../../../manifests/models/glm-5v-turbo.json' import Gpt41 from '../../../manifests/models/gpt-4-1.json' @@ -194,6 +195,7 @@ export const modelsData = [ Glm47, Glm51, Glm52, + Glm53, Glm5Turbo, Glm5, Glm5vTurbo, diff --git a/tests/model-intelligence-index.test.ts b/tests/model-intelligence-index.test.ts index 48555bcf..27a7eaea 100644 --- a/tests/model-intelligence-index.test.ts +++ b/tests/model-intelligence-index.test.ts @@ -499,6 +499,7 @@ describe('model intelligence index', () => { ['glm-5v-turbo', 'GLM Vision', 35], ['glm-5-1', 'GLM', 41], ['glm-5-2', 'GLM', 53], + ['glm-5-3', 'GLM', 60], ]) }) diff --git a/tests/model-price-intelligence-index.test.ts b/tests/model-price-intelligence-index.test.ts index 5e5fa6e5..8a8f23d5 100644 --- a/tests/model-price-intelligence-index.test.ts +++ b/tests/model-price-intelligence-index.test.ts @@ -19,6 +19,7 @@ describe('model price-intelligence index', () => { 'gemini-3-7-flash', 'qwen3-7-plus', 'qwen3-8-2-4t-a95b', + 'glm-5-3', 'gpt-5-6-terra', 'claude-haiku-4-5', 'claude-opus-5', @@ -27,7 +28,7 @@ describe('model price-intelligence index', () => { ]) ) expect(modelPriceIntelligencePoints.map(point => point.modelId)).not.toEqual( - expect.arrayContaining(['claude-opus-4-8', 'muse-spark-1-1', 'mistral-medium-3-5']) + expect.arrayContaining(['claude-opus-4-8', 'glm-5-2', 'muse-spark-1-1', 'mistral-medium-3-5']) ) })