From fd1b11cfc765fc6f17af7085eb8915ef7d33cf17 Mon Sep 17 00:00:00 2001 From: phumchai1515-prog Date: Wed, 12 Aug 2026 06:36:48 +0700 Subject: [PATCH] docs: refresh model-selection guidance to the Claude 5 families (#2723) Model-routing guidance across the rules, skills, and harness-steering docs still recommends Sonnet 4.6 / Opus 4.5-4.6 by name. Readers on the current generation have to map those onto Sonnet 5 / Opus 5 themselves, and the recommendation reads as pinned to a superseded generation. Renames the recommended models in guidance tables and updates two pinned model IDs in code samples: - rules/steering guidance: .cursor, .kiro, and the seven translated performance.md copies (ja-JP, zh-CN, zh-TW, ko-KR, pt-BR, es, tr) - skills/prompt-optimizer complexity-routing table (+ zh-CN copy) - skills/cost-aware-llm-pipeline MODEL_SONNET constant (+ zh-CN, ja-JP) - docs/examples project-guidelines template, which pinned the invalid ID claude-sonnet-4-5-20250514 (+ zh-TW, ja-JP copies) Deliberately left alone: - The "Pricing Reference (2025-2026)" table in cost-aware-llm-pipeline. Renaming those rows while keeping the existing per-token figures would assert Claude 5 pricing this change has not verified. - Executable model config (.opencode/opencode.json, agent.yaml). Those pins change real agent behavior and belong in their own reviewed change. - Historical and illustrative references: the-shortform-guide session transcripts, the ECC-PRO roadmap log entry, gan-style-harness's "Opus 4.5-class"/"Opus 4.6-class" capability tiers, and strategic-compact's deliberately generic "400k Opus 4.x" example. - docs/ATLAS-CLOUD-GUIDE.md, which lists a third-party provider's catalog. Documentation wording only; no behavioral change. Co-authored-by: Phumchai Tanonsi <274848436+phumchai1515-prog@users.noreply.github.com> --- .cursor/rules/common-performance.md | 4 ++-- .kiro/steering/performance.md | 4 ++-- docs/es/rules/common/performance.md | 4 ++-- docs/examples/project-guidelines-template.md | 2 +- docs/ja-JP/rules/common/performance.md | 4 ++-- docs/ja-JP/skills/cost-aware-llm-pipeline/SKILL.md | 2 +- .../ja-JP/skills/project-guidelines-example/SKILL.md | 2 +- docs/ko-KR/rules/performance.md | 4 ++-- docs/pt-BR/rules/performance.md | 4 ++-- docs/tr/rules/common/performance.md | 4 ++-- docs/zh-CN/rules/common/performance.md | 4 ++-- docs/zh-CN/skills/cost-aware-llm-pipeline/SKILL.md | 2 +- docs/zh-CN/skills/prompt-optimizer/SKILL.md | 12 ++++++------ docs/zh-TW/rules/performance.md | 4 ++-- .../zh-TW/skills/project-guidelines-example/SKILL.md | 2 +- skills/cost-aware-llm-pipeline/SKILL.md | 2 +- skills/prompt-optimizer/SKILL.md | 12 ++++++------ 17 files changed, 36 insertions(+), 36 deletions(-) diff --git a/.cursor/rules/common-performance.md b/.cursor/rules/common-performance.md index ec0f93a29..ef4114025 100644 --- a/.cursor/rules/common-performance.md +++ b/.cursor/rules/common-performance.md @@ -11,12 +11,12 @@ alwaysApply: true - Pair programming and code generation - Worker agents in multi-agent systems -**Sonnet 4.6** (Best coding model): +**Sonnet 5** (Best coding model): - Main development work - Orchestrating multi-agent workflows - Complex coding tasks -**Opus 4.6** (Deepest reasoning): +**Opus 5** (Deepest reasoning): - Complex architectural decisions - Maximum reasoning requirements - Research and analysis tasks diff --git a/.kiro/steering/performance.md b/.kiro/steering/performance.md index a5638a34d..7cae57041 100644 --- a/.kiro/steering/performance.md +++ b/.kiro/steering/performance.md @@ -13,12 +13,12 @@ description: Performance optimization guidelines including model selection strat - Pair programming and code generation - Worker agents in multi-agent systems -**Claude Sonnet 4.6** (Best coding model): +**Claude Sonnet 5** (Best coding model): - Main development work - Orchestrating multi-agent workflows - Complex coding tasks -**Claude Opus 4.6** (Deepest reasoning): +**Claude Opus 5** (Deepest reasoning): - Complex architectural decisions - Maximum reasoning requirements - Research and analysis tasks diff --git a/docs/es/rules/common/performance.md b/docs/es/rules/common/performance.md index 53ddcead6..6f4dda544 100644 --- a/docs/es/rules/common/performance.md +++ b/docs/es/rules/common/performance.md @@ -7,12 +7,12 @@ - Programación en pareja y generación de código - Agentes workers en sistemas multi-agente -**Sonnet 4.6** (Mejor modelo para codificación): +**Sonnet 5** (Mejor modelo para codificación): - Trabajo de desarrollo principal - Orquestación de flujos de trabajo multi-agente - Tareas de codificación complejas -**Opus 4.5** (Razonamiento más profundo): +**Opus 5** (Razonamiento más profundo): - Decisiones arquitectónicas complejas - Requisitos de razonamiento máximo - Tareas de investigación y análisis diff --git a/docs/examples/project-guidelines-template.md b/docs/examples/project-guidelines-template.md index b3e7fc73a..8290905ea 100644 --- a/docs/examples/project-guidelines-template.md +++ b/docs/examples/project-guidelines-template.md @@ -161,7 +161,7 @@ async def analyze_with_claude(content: str) -> AnalysisResult: client = Anthropic() response = client.messages.create( - model="claude-sonnet-4-5-20250514", + model="claude-sonnet-5", max_tokens=1024, messages=[{"role": "user", "content": content}], tools=[{ diff --git a/docs/ja-JP/rules/common/performance.md b/docs/ja-JP/rules/common/performance.md index c97ed1b03..f08ba0c4e 100644 --- a/docs/ja-JP/rules/common/performance.md +++ b/docs/ja-JP/rules/common/performance.md @@ -7,12 +7,12 @@ - ペアプログラミングとコード生成 - マルチ agent システムのワーカー agent -**Sonnet 4.6**（最高のコーディングモデル）: +**Sonnet 5**（最高のコーディングモデル）: - メイン開発作業 - マルチ agent ワークフローのオーケストレーション - 複雑なコーディングタスク -**Opus 4.6**（最も深い推論）: +**Opus 5**（最も深い推論）: - 複雑なアーキテクチャの意思決定 - 最大限の推論要件 - 調査と分析タスク diff --git a/docs/ja-JP/skills/cost-aware-llm-pipeline/SKILL.md b/docs/ja-JP/skills/cost-aware-llm-pipeline/SKILL.md index 97e95f06d..0759a3a62 100644 --- a/docs/ja-JP/skills/cost-aware-llm-pipeline/SKILL.md +++ b/docs/ja-JP/skills/cost-aware-llm-pipeline/SKILL.md @@ -22,7 +22,7 @@ origin: ECC シンプルなタスクには自動的に安価なモデルを選択し、複雑なタスクのために高価なモデルを予約します。 ```python -MODEL_SONNET = "claude-sonnet-4-6" +MODEL_SONNET = "claude-sonnet-5" MODEL_HAIKU = "claude-haiku-4-5-20251001" _SONNET_TEXT_THRESHOLD = 10_000 # 文字数 diff --git a/docs/ja-JP/skills/project-guidelines-example/SKILL.md b/docs/ja-JP/skills/project-guidelines-example/SKILL.md index 4e4f06731..90dde10c6 100644 --- a/docs/ja-JP/skills/project-guidelines-example/SKILL.md +++ b/docs/ja-JP/skills/project-guidelines-example/SKILL.md @@ -166,7 +166,7 @@ async def analyze_with_claude(content: str) -> AnalysisResult: client = Anthropic() response = client.messages.create( - model="claude-sonnet-4-5-20250514", + model="claude-sonnet-5", max_tokens=1024, messages=[{"role": "user", "content": content}], tools=[{ diff --git a/docs/ko-KR/rules/performance.md b/docs/ko-KR/rules/performance.md index 931925b6a..efb85402a 100644 --- a/docs/ko-KR/rules/performance.md +++ b/docs/ko-KR/rules/performance.md @@ -7,12 +7,12 @@ - 페어 프로그래밍과 코드 생성 - 멀티 에이전트 시스템의 워커 에이전트 -**Sonnet 4.6** (최고의 코딩 모델): +**Sonnet 5** (최고의 코딩 모델): - 주요 개발 작업 - 멀티 에이전트 워크플로우 오케스트레이션 - 복잡한 코딩 작업 -**Opus 4.6** (가장 깊은 추론): +**Opus 5** (가장 깊은 추론): - 복잡한 아키텍처 의사결정 - 최대 추론 요구사항 - 리서치 및 분석 작업 diff --git a/docs/pt-BR/rules/performance.md b/docs/pt-BR/rules/performance.md index 07f5cd342..696eba888 100644 --- a/docs/pt-BR/rules/performance.md +++ b/docs/pt-BR/rules/performance.md @@ -7,12 +7,12 @@ - Programação em par e geração de código - Agentes worker em sistemas multi-agente -**Sonnet 4.6** (Melhor modelo para codificação): +**Sonnet 5** (Melhor modelo para codificação): - Trabalho principal de desenvolvimento - Orquestrando fluxos de trabalho multi-agente - Tarefas de codificação complexas -**Opus 4.6** (Raciocínio mais profundo): +**Opus 5** (Raciocínio mais profundo): - Decisões arquiteturais complexas - Requisitos máximos de raciocínio - Pesquisa e análise diff --git a/docs/tr/rules/common/performance.md b/docs/tr/rules/common/performance.md index 2312099ba..72166c13a 100644 --- a/docs/tr/rules/common/performance.md +++ b/docs/tr/rules/common/performance.md @@ -7,12 +7,12 @@ - Pair programming ve kod üretimi - Multi-agent sistemlerinde worker agent'lar -**Sonnet 4.6** (En iyi kodlama modeli): +**Sonnet 5** (En iyi kodlama modeli): - Ana geliştirme çalışması - Multi-agent iş akışlarını orkestrasyon - Karmaşık kodlama görevleri -**Opus 4.6** (En derin akıl yürütme): +**Opus 5** (En derin akıl yürütme): - Karmaşık mimari kararlar - Maksimum akıl yürütme gereksinimleri - Araştırma ve analiz görevleri diff --git a/docs/zh-CN/rules/common/performance.md b/docs/zh-CN/rules/common/performance.md index 7dcf4d895..9c1753686 100644 --- a/docs/zh-CN/rules/common/performance.md +++ b/docs/zh-CN/rules/common/performance.md @@ -8,13 +8,13 @@ * 结对编程和代码生成 * 多智能体系统中的工作智能体 -**Sonnet 4.6** (最佳编码模型): +**Sonnet 5** (最佳编码模型): * 主要的开发工作 * 编排多智能体工作流 * 复杂的编码任务 -**Opus 4.6** (最深的推理能力): +**Opus 5** (最深的推理能力): * 复杂的架构决策 * 最高级别的推理需求 diff --git a/docs/zh-CN/skills/cost-aware-llm-pipeline/SKILL.md b/docs/zh-CN/skills/cost-aware-llm-pipeline/SKILL.md index 9af5a8466..869deec11 100644 --- a/docs/zh-CN/skills/cost-aware-llm-pipeline/SKILL.md +++ b/docs/zh-CN/skills/cost-aware-llm-pipeline/SKILL.md @@ -22,7 +22,7 @@ origin: ECC 自动为简单任务选择更便宜的模型，为复杂任务保留昂贵的模型。 ```python -MODEL_SONNET = "claude-sonnet-4-6" +MODEL_SONNET = "claude-sonnet-5" MODEL_HAIKU = "claude-haiku-4-5-20251001" _SONNET_TEXT_THRESHOLD = 10_000 # chars diff --git a/docs/zh-CN/skills/prompt-optimizer/SKILL.md b/docs/zh-CN/skills/prompt-optimizer/SKILL.md index d833aec2a..76edd7cfc 100644 --- a/docs/zh-CN/skills/prompt-optimizer/SKILL.md +++ b/docs/zh-CN/skills/prompt-optimizer/SKILL.md @@ -158,10 +158,10 @@ Research → Plan → Implement (TDD) → Review → Verify → Commit | 范围 | 推荐模型 | 理由 | |-------|------------------|-----------| -| 微小-低 | Sonnet 4.6 | 快速、成本效益高，适合简单任务 | -| 中 | Sonnet 4.6 | 标准工作的最佳编码模型 | -| 高 | Sonnet 4.6 (主) + Opus 4.6 (规划) | Opus 用于架构，Sonnet 用于实现 | -| 史诗级 | Opus 4.6 (蓝图) + Sonnet 4.6 (执行) | 深度推理用于多会话规划 | +| 微小-低 | Sonnet 5 | 快速、成本效益高，适合简单任务 | +| 中 | Sonnet 5 | 标准工作的最佳编码模型 | +| 高 | Sonnet 5 (主) + Opus 5 (规划) | Opus 用于架构，Sonnet 用于实现 | +| 史诗级 | Opus 5 (蓝图) + Sonnet 5 (执行) | 深度推理用于多会话规划 | **多提示拆分**（针对高/史诗级范围）： @@ -197,7 +197,7 @@ Research → Plan → Implement (TDD) → Review → Verify → Commit | 命令 | /plan | 编码前规划架构 | | 技能 | tdd-workflow | TDD 方法指导 | | 代理 | code-reviewer | 实施后审查 | -| 模型 | Sonnet 4.6 | 针对此范围的推荐模型 | +| 模型 | Sonnet 5 | 针对此范围的推荐模型 | ### 第 3 部分：优化提示 —— 完整版本 @@ -363,7 +363,7 @@ Research → Plan → Implement (TDD) → Review → Verify → Commit 阶段之间使用 /save-session。使用 /resume-session 继续。 在依赖关系允许时，使用 git worktrees 进行并行服务提取。 -推荐：使用 Opus 4.6 进行蓝图规划，使用 Sonnet 4.6 执行各阶段。 +推荐：使用 Opus 5 进行蓝图规划，使用 Sonnet 5 执行各阶段。 ``` *** diff --git a/docs/zh-TW/rules/performance.md b/docs/zh-TW/rules/performance.md index 78f85c6b7..f001bc72b 100644 --- a/docs/zh-TW/rules/performance.md +++ b/docs/zh-TW/rules/performance.md @@ -7,12 +7,12 @@ - 配對程式設計和程式碼產生 - 多 agent 系統中的 worker agents -**Sonnet 4.6**（最佳程式碼模型）： +**Sonnet 5**（最佳程式碼模型）： - 主要開發工作 - 協調多 agent 工作流程 - 複雜程式碼任務 -**Opus 4.6**（最深度推理）： +**Opus 5**（最深度推理）： - 複雜架構決策 - 最大推理需求 - 研究和分析任務 diff --git a/docs/zh-TW/skills/project-guidelines-example/SKILL.md b/docs/zh-TW/skills/project-guidelines-example/SKILL.md index acc690c8e..4e7f55084 100644 --- a/docs/zh-TW/skills/project-guidelines-example/SKILL.md +++ b/docs/zh-TW/skills/project-guidelines-example/SKILL.md @@ -166,7 +166,7 @@ async def analyze_with_claude(content: str) -> AnalysisResult: client = Anthropic() response = client.messages.create( - model="claude-sonnet-4-5-20250514", + model="claude-sonnet-5", max_tokens=1024, messages=[{"role": "user", "content": content}], tools=[{ diff --git a/skills/cost-aware-llm-pipeline/SKILL.md b/skills/cost-aware-llm-pipeline/SKILL.md index 139d10985..590a7ecdd 100644 --- a/skills/cost-aware-llm-pipeline/SKILL.md +++ b/skills/cost-aware-llm-pipeline/SKILL.md @@ -23,7 +23,7 @@ Patterns for controlling LLM API costs while maintaining quality. Combines model Automatically select cheaper models for simple tasks, reserving expensive models for complex ones. ```python -MODEL_SONNET = "claude-sonnet-4-6" +MODEL_SONNET = "claude-sonnet-5" MODEL_HAIKU = "claude-haiku-4-5-20251001" _SONNET_TEXT_THRESHOLD = 10_000 # chars diff --git a/skills/prompt-optimizer/SKILL.md b/skills/prompt-optimizer/SKILL.md index 6a7a2fed1..0d486bac4 100644 --- a/skills/prompt-optimizer/SKILL.md +++ b/skills/prompt-optimizer/SKILL.md @@ -179,10 +179,10 @@ For MEDIUM+ tasks, always start with /plan. For EPIC tasks, use blueprint skill. | Scope | Recommended Model | Rationale | |-------|------------------|-----------| -| TRIVIAL-LOW | Sonnet 4.6 | Fast, cost-efficient for simple tasks | -| MEDIUM | Sonnet 4.6 | Best coding model for standard work | -| HIGH | Sonnet 4.6 (main) + Opus 4.6 (planning) | Opus for architecture, Sonnet for implementation | -| EPIC | Opus 4.6 (blueprint) + Sonnet 4.6 (execution) | Deep reasoning for multi-session planning | +| TRIVIAL-LOW | Sonnet 5 | Fast, cost-efficient for simple tasks | +| MEDIUM | Sonnet 5 | Best coding model for standard work | +| HIGH | Sonnet 5 (main) + Opus 5 (planning) | Opus for architecture, Sonnet for implementation | +| EPIC | Opus 5 (blueprint) + Sonnet 5 (execution) | Deep reasoning for multi-session planning | **Multi-prompt splitting** (for HIGH/EPIC scope): @@ -219,7 +219,7 @@ If Phase 0 auto-detected the answer, state it instead of asking. | Command | /plan | Plan architecture before coding | | Skill | tdd-workflow | TDD methodology guidance | | Agent | code-reviewer | Post-implementation review | -| Model | Sonnet 4.6 | Recommended for this scope | +| Model | Sonnet 5 | Recommended for this scope | ### Section 3: Optimized Prompt — Full Version @@ -381,7 +381,7 @@ Each phase = 1 PR, with /verify gates between phases. Use /save-session between phases. Use /resume-session to continue. Use git worktrees for parallel service extraction when dependencies allow. -Recommended: Opus 4.6 for blueprint planning, Sonnet 4.6 for phase execution. +Recommended: Opus 5 for blueprint planning, Sonnet 5 for phase execution. ``` ---