2026-06-16 11:59:25 +02:00
---
created: 2026-06-16
2026-07-07 10:31:37 +02:00
updated: 2026-07-07
sources: [other/2026-06-16_openclaw-releases-v2026.6.8.md, youtube/2026-07-07_berman-model-routing.md]
tags: [architecture, model-routing, llm, openclaw, planning-execution-split, cost-savings]
2026-06-16 11:59:25 +02:00
---
2026-06-07 11:47:41 +02:00
# Model Routing
> Migriert aus ByteRover context-tree (April 2026)
## Two-Model-Pipeline
Für mehrstufige Aufgaben: **Grok 4.20 → Gemini 2.5 Flash**
| Phase | Modell | Stärke |
|-------|--------|--------|
| Ideation | Grok 4.20 | Kreative Exploration, Brainstorming, divergentes Denken |
| Synthesis | Gemini 2.5 Flash | Strukturierte Ausgabe, schnelle Synthese, konvergent |
Dieses Pattern wird sowohl im allgemeinen Routing als auch im Subconscious Agent (Hard Synthesis) verwendet, aber nie als Shared Primitive formalisiert.
## GPT-5.4 Konfiguration
- **Alias:** `GPT-5.4`
- **Route:** `openrouter/openai/gpt-5.4`
- **Context:** 200k Tokens
- **Output:** 32k Tokens
- **Reasoning/Thinking:** ❌ Deaktiviert (per Claire Vo-Empfehlung)
## Fallback-Chain (Hector Primary)
```
openrouter/auto
→ zai/glm-5-turbo
→ moonshot/moonshot-v1-auto
→ moonshot/kimi-k2.5
→ ollama/minimax-m2.7
2026-06-15 09:19:22 +02:00
```
ingest(github): openclaw releases v2026.6.6-v2026.6.8
Raw dump of top-5 OpenClaw releases (16.06.2026) via GitHub REST API.
Highlights of the digest:
- v2026.6.8 stable: GLM-5.2 + Claude Haiku 4.5 catalog rows, OpenRouter/Vertex
provider-prefix normalization, Telegram rich-message delivery, QMD memory
search stable in transient mode, raw Memory Wiki source pages as source
evidence, Skill Workshop symlink-write validation
- v2026.6.7-beta.1: Kimi K2.7 Code added, Discord thread-title timeout,
Anthropic thinking replay repair
- v2026.6.6 stable: major security release — exec approvals fail closed on
timeout, MCP stdio hardening, Telegram account-scoped topics
Wiki changes:
- tools/openclaw.md (new): platform reference with version timeline, 10
plattformrelevante changes vs our setup, rollout plan (awaiting Kai approval
for v2026.6.8 update)
- architecture/model-routing.md (updated): GLM-5.2-Support + provider-prefix
normalization for v2026.6.8, distribution matrix updated
- architecture/memory-system.md (updated): QMD stabilization, SQLite-NFS-WAL
detection, raw-Memory-Wiki-source-pages as source evidence, full-reindex
rollback recovery
- concepts/llm/glm-5.2-zai-coding-model.md (updated): OpenClaw v2026.6.8
native provider adapter breaks Z.ai Coding Plan monopoly
- index.md (updated): new tools entry, header datum 12. update, updated
descriptions for memory/routing/glm-5.2, new raw sources row
- log.md (updated): new ingest entry following schema
2026-06-16 20:17:21 +02:00
**Hinweis (Stand 2026-06-16):** Wir sind gerade dabei, die Fallback-Chain
auf ** `ollama/glm-5.1` als Primary + `ollama/minimax-m3` als Fallback**
umzustellen. Details unten in "Update 2026-06-16: OpenClaw v2026.6.8".
## Update 2026-06-16: OpenClaw v2026.6.8 GLM-5.2-Support + Provider-Prefix-Normalisierung
Mit dem Release [v2026.6.8 ](https://github.com/openclaw/openclaw/releases/tag/v2026.6.8 )
(16.06.2026) hat OpenClaw **drei für uns routing-relevante Änderungen** gebracht:
1. **GLM-5.2 ist offiziell im Provider-Catalog** — vorher nur über Z.ai Coding
Plan verfügbar, jetzt nativ in OpenClaw routeable. Siehe
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
[[../concepts/llm/glm-5.2-zai-coding-model.md]].
ingest(github): openclaw releases v2026.6.6-v2026.6.8
Raw dump of top-5 OpenClaw releases (16.06.2026) via GitHub REST API.
Highlights of the digest:
- v2026.6.8 stable: GLM-5.2 + Claude Haiku 4.5 catalog rows, OpenRouter/Vertex
provider-prefix normalization, Telegram rich-message delivery, QMD memory
search stable in transient mode, raw Memory Wiki source pages as source
evidence, Skill Workshop symlink-write validation
- v2026.6.7-beta.1: Kimi K2.7 Code added, Discord thread-title timeout,
Anthropic thinking replay repair
- v2026.6.6 stable: major security release — exec approvals fail closed on
timeout, MCP stdio hardening, Telegram account-scoped topics
Wiki changes:
- tools/openclaw.md (new): platform reference with version timeline, 10
plattformrelevante changes vs our setup, rollout plan (awaiting Kai approval
for v2026.6.8 update)
- architecture/model-routing.md (updated): GLM-5.2-Support + provider-prefix
normalization for v2026.6.8, distribution matrix updated
- architecture/memory-system.md (updated): QMD stabilization, SQLite-NFS-WAL
detection, raw-Memory-Wiki-source-pages as source evidence, full-reindex
rollback recovery
- concepts/llm/glm-5.2-zai-coding-model.md (updated): OpenClaw v2026.6.8
native provider adapter breaks Z.ai Coding Plan monopoly
- index.md (updated): new tools entry, header datum 12. update, updated
descriptions for memory/routing/glm-5.2, new raw sources row
- log.md (updated): new ingest entry following schema
2026-06-16 20:17:21 +02:00
2. **Provider-Prefix-Normalisierung** für OpenRouter und Google Vertex Pfade:
`openrouter/anthropic/claude-3.5-sonnet` wird intern zu bare IDs
normalisiert wo Runtimes das brauchen.
3. **OAuth image-default routing durch Codex** für eligible OpenAI profiles —
relevant für Generated-Media-Workflows (image_generate, music_generate).
**Konkrete Implikation für unsere aktuelle Konfiguration:**
| Setting | Status | Aktion |
|---|---|---|
| `ollama/minimax-m3` (Primary, lokal) | ✅ Unverändert | keine |
| `ollama/glm-5.1` (geplant als neuer Primary) | ⚠️ Migration offen | siehe unten |
| `openrouter/auto` (Fallback) | ✅ Unverändert | keine — Prefix-Norm ist transparent |
| GLM-5.2 als Sub-Route | 🆕 Option | nach v2026.6.8-Update evaluieren |
**Aktueller Routing-Plan nach v2026.6.8-Migration:**
```
Primary: ollama/glm-5.1
Fallback: ollama/minimax-m3
Optional (nach Evaluation):
- openrouter/anthropic/claude-sonnet-4.5 (Haiku 4.5 catalog row in v2026.6.8)
- zai/glm-5.2 (Long-Context, 1M)
- moonshot/kimi-k2.7-code (Coding, agentic)
```
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
Vollständige Plattform-Doku: [[../tools/openclaw.md]].
ingest(github): openclaw releases v2026.6.6-v2026.6.8
Raw dump of top-5 OpenClaw releases (16.06.2026) via GitHub REST API.
Highlights of the digest:
- v2026.6.8 stable: GLM-5.2 + Claude Haiku 4.5 catalog rows, OpenRouter/Vertex
provider-prefix normalization, Telegram rich-message delivery, QMD memory
search stable in transient mode, raw Memory Wiki source pages as source
evidence, Skill Workshop symlink-write validation
- v2026.6.7-beta.1: Kimi K2.7 Code added, Discord thread-title timeout,
Anthropic thinking replay repair
- v2026.6.6 stable: major security release — exec approvals fail closed on
timeout, MCP stdio hardening, Telegram account-scoped topics
Wiki changes:
- tools/openclaw.md (new): platform reference with version timeline, 10
plattformrelevante changes vs our setup, rollout plan (awaiting Kai approval
for v2026.6.8 update)
- architecture/model-routing.md (updated): GLM-5.2-Support + provider-prefix
normalization for v2026.6.8, distribution matrix updated
- architecture/memory-system.md (updated): QMD stabilization, SQLite-NFS-WAL
detection, raw-Memory-Wiki-source-pages as source evidence, full-reindex
rollback recovery
- concepts/llm/glm-5.2-zai-coding-model.md (updated): OpenClaw v2026.6.8
native provider adapter breaks Z.ai Coding Plan monopoly
- index.md (updated): new tools entry, header datum 12. update, updated
descriptions for memory/routing/glm-5.2, new raw sources row
- log.md (updated): new ingest entry following schema
2026-06-16 20:17:21 +02:00
2026-06-15 09:19:22 +02:00
## Verwandt: Parallele Ensembles (Fusion)
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
Sequentielles Routing (Fallback) ist nicht das einzige Pattern. OpenRouter bietet **Fusion** — parallele Model-Panels mit Judge-Synthese. Siehe [[../concepts/llm/llm-model-fusion-ensembles.md]] für Details und DRACO-Benchmark-Ergebnisse.
2026-06-15 09:19:22 +02:00
ingest(youtube): fahd-mirza-kimi-k2.7-vs-glm-5.2
Fahd Mirza Head-to-Head in Hermes Agent (14.06.2026, 14 Min, 2659 views,
89 likes). Beide Modelle bekommen denselben Prompt, loesen denselben
Real-World-Task: Flask-App mit gepflanztem FIFA-Bug + Round-of-32-
Bracket-Feature mit Anti-Group-Rematch-Regel.
Test 1: Beide bestehen. Kimi schneller (~5 Min), innovativer
(zusaetzliche Progression-Previews). GLM: 97 tool calls, laenger.
Test 2 (Creative HTML, Siberian Wind): GLM staerker bei Animation +
Terrain-Detail; Kimi bei Stats + Geographie.
Konzept: Real-World-Showdown als Benchmark-Alternative zu statischen
Tests (HumanEval, MBPP, SWE-bench). Agent-Framework isoliert die
Modell-Variable sauber.
Direkte Implikation fuer OpenClaw:
- Sub-Task-spezifisches Coding-Routing (Kimi vs GLM je nach Task)
- Ensemble-Kandidaten: Kimi K2.7 + GLM-5.2 in Coding-Panels
- Real-World-Validierung statt nur Standard-Benchmarks fuer Production-
Pfade
- Hermes Agent als alternatives Agent-Framework zu AutoGen/CrewAI
Aenderungen:
- raw/youtube/2026-06-14_fahd-mirza-kimi-k2.7-vs-glm-5.2.md (neu)
- wiki/concepts/real-world-coding-showdown.md (neu)
- wiki/tools/kimi-k2.7-code.md (Cross-References)
- wiki/architecture/model-routing.md (neue Sektion Coding-Modelle)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung gemaess AGENTS.md-Kardinalregel:
- 12 externe Verweise auf Modelle (Kimi, GLM, Hermes, Ollama, Z.ai)
- 9 Benchmark-/Methoden-Links (SWE-bench, HumanEval, MBPP, MCP, etc.)
- 7 Fahd-Mirza-Links (YT, Blog, LinkedIn, Substack, Ko-Fi)
- 5 Wiki-Cross-References (fusion, ai-agents, kimi-k2.7, model-routing, behavior-persistence)
2026-06-15 09:45:38 +02:00
**Aktuell offene Frage:** Soll OpenClaws Quality-kritische Pfade (z.B. Subconscious Hard Synthesis) um eine parallele Pre-Routing-Schicht erweitert werden? Trade-off: höhere Latenz + Kosten vs. bessere Qualität durch Modell-Heterogenität.
## Coding-Modelle: Kimi K2.7 & GLM-5.2
Im Juni 2026 haben sich zwei starke Open-Source-Coding-Modelle aus China etabliert:
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
- **Kimi K2.7** (Moonshot AI) — siehe [[../tools/kimi-k2.7-code.md]]. MoE 1T/32B, 256K Kontext, schnell im Real-Coding-Showdown.
ingest(youtube): fahd-mirza-kimi-k2.7-vs-glm-5.2
Fahd Mirza Head-to-Head in Hermes Agent (14.06.2026, 14 Min, 2659 views,
89 likes). Beide Modelle bekommen denselben Prompt, loesen denselben
Real-World-Task: Flask-App mit gepflanztem FIFA-Bug + Round-of-32-
Bracket-Feature mit Anti-Group-Rematch-Regel.
Test 1: Beide bestehen. Kimi schneller (~5 Min), innovativer
(zusaetzliche Progression-Previews). GLM: 97 tool calls, laenger.
Test 2 (Creative HTML, Siberian Wind): GLM staerker bei Animation +
Terrain-Detail; Kimi bei Stats + Geographie.
Konzept: Real-World-Showdown als Benchmark-Alternative zu statischen
Tests (HumanEval, MBPP, SWE-bench). Agent-Framework isoliert die
Modell-Variable sauber.
Direkte Implikation fuer OpenClaw:
- Sub-Task-spezifisches Coding-Routing (Kimi vs GLM je nach Task)
- Ensemble-Kandidaten: Kimi K2.7 + GLM-5.2 in Coding-Panels
- Real-World-Validierung statt nur Standard-Benchmarks fuer Production-
Pfade
- Hermes Agent als alternatives Agent-Framework zu AutoGen/CrewAI
Aenderungen:
- raw/youtube/2026-06-14_fahd-mirza-kimi-k2.7-vs-glm-5.2.md (neu)
- wiki/concepts/real-world-coding-showdown.md (neu)
- wiki/tools/kimi-k2.7-code.md (Cross-References)
- wiki/architecture/model-routing.md (neue Sektion Coding-Modelle)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung gemaess AGENTS.md-Kardinalregel:
- 12 externe Verweise auf Modelle (Kimi, GLM, Hermes, Ollama, Z.ai)
- 9 Benchmark-/Methoden-Links (SWE-bench, HumanEval, MBPP, MCP, etc.)
- 7 Fahd-Mirza-Links (YT, Blog, LinkedIn, Substack, Ko-Fi)
- 5 Wiki-Cross-References (fusion, ai-agents, kimi-k2.7, model-routing, behavior-persistence)
2026-06-15 09:45:38 +02:00
- **GLM-5.2** (Zhipu AI / Z.ai) — 744B Dense, 1M Kontext, MIT-Lizenz, stark bei kreativer Generation und Animation.
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
Head-to-Head-Vergleich in Hermes Agent: siehe [[../concepts/llm/real-world-coding-showdown.md]].
ingest(youtube): fahd-mirza-kimi-k2.7-vs-glm-5.2
Fahd Mirza Head-to-Head in Hermes Agent (14.06.2026, 14 Min, 2659 views,
89 likes). Beide Modelle bekommen denselben Prompt, loesen denselben
Real-World-Task: Flask-App mit gepflanztem FIFA-Bug + Round-of-32-
Bracket-Feature mit Anti-Group-Rematch-Regel.
Test 1: Beide bestehen. Kimi schneller (~5 Min), innovativer
(zusaetzliche Progression-Previews). GLM: 97 tool calls, laenger.
Test 2 (Creative HTML, Siberian Wind): GLM staerker bei Animation +
Terrain-Detail; Kimi bei Stats + Geographie.
Konzept: Real-World-Showdown als Benchmark-Alternative zu statischen
Tests (HumanEval, MBPP, SWE-bench). Agent-Framework isoliert die
Modell-Variable sauber.
Direkte Implikation fuer OpenClaw:
- Sub-Task-spezifisches Coding-Routing (Kimi vs GLM je nach Task)
- Ensemble-Kandidaten: Kimi K2.7 + GLM-5.2 in Coding-Panels
- Real-World-Validierung statt nur Standard-Benchmarks fuer Production-
Pfade
- Hermes Agent als alternatives Agent-Framework zu AutoGen/CrewAI
Aenderungen:
- raw/youtube/2026-06-14_fahd-mirza-kimi-k2.7-vs-glm-5.2.md (neu)
- wiki/concepts/real-world-coding-showdown.md (neu)
- wiki/tools/kimi-k2.7-code.md (Cross-References)
- wiki/architecture/model-routing.md (neue Sektion Coding-Modelle)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung gemaess AGENTS.md-Kardinalregel:
- 12 externe Verweise auf Modelle (Kimi, GLM, Hermes, Ollama, Z.ai)
- 9 Benchmark-/Methoden-Links (SWE-bench, HumanEval, MBPP, MCP, etc.)
- 7 Fahd-Mirza-Links (YT, Blog, LinkedIn, Substack, Ko-Fi)
- 5 Wiki-Cross-References (fusion, ai-agents, kimi-k2.7, model-routing, behavior-persistence)
2026-06-15 09:45:38 +02:00
ingest(youtube): ichbinfabian-glm-5.2
GLM 5.2 Deep-Dive (IAmFabian, 14.06.2026, 13:15 Min, 212 views,
16 Likes). Deutscher Reviewer testet live: Flappy Bird (1:20, 481
Zeilen HTML, laeuft) und Newton-Pendel mit echter Physik (9:41, davon
8:27 nur Thinking).
Kernkonzepte:
- Chinas Frontier-Coding-Modelle 2026: Kimi K2.7 + GLM 5.2, beide
Frontier-Niveau bei 10x niedrigeren Preisen
- GLM 5.2: 1M Kontext, MIT-Lizenz (Open Weights 'naechste Woche'),
laut AI Code King Benchmarks nur 6% unter Opus 4.8
- Z.ai Coding Plan (18 USD/Monat Light) als exklusive Distribution
- Light-Plan: 2-3x Kontingent-Mehrverbrauch in Stosszeiten
- OpenRouter-Eintrag 'in den naechsten Tagen' erwartet
- Coding Helper (npx @z_ai/coding-helper) bindet GLM 5.2 in Claude
Code, OpenCode, Cline, Kilo Code, Crush, Factory Droid ein
KRITISCHER IMPLIKAT FUER OPENCLAW:
- Z.ai Coding Plan listet 'Clawdbot/OpenClaw' naamentlich als
unterstuetztes Environment - direkte Integration bereits vorbereitet
- Drei Setup-Optionen:
1. Z.ai Coding Plan sofort (18 USD/Monat, Light)
2. Self-Hosting nach Open Weights (MIT, HuggingFace)
3. Warten auf OpenRouter (Tage)
Aenderungen:
- raw/youtube/2026-06-15_ichbinfabian-glm-5.2-coding-modell.md (neu)
- wiki/concepts/glm-5.2-zai-coding-model.md (neu)
- wiki/architecture/model-routing.md (neue Sektion GLM 5.2 +
Routing-Position + 3 Setup-Optionen)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung: 12 Z.ai/HuggingFace/NPM-Links, 10 Tool-Links
(OpenCode, Claude Code, Cline, Kilo Code, Crush, Droid),
8 Wiki-Cross-References, 4 Konkurrenz-Links (Kimi, OpenRouter,
Open LLM Leaderboard, Anthropic).
2026-06-15 09:50:23 +02:00
**Implikation für Routing:** In Sub-Task-Spezialisierung (Coding vs. Creative vs. Long-Context) sind beide als Sub-Routing-Knoten sinnvoll — nicht als Ersatz der Primary-Fallback-Chain, sondern als **spezialisierte Side-Branches** für Coding-Tasks.
### GLM 5.2 (Z.ai) — 1M Kontext, MIT-Lizenz, 10x günstiger
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
Zusätzlich zu Kimi K2.7 ist **GLM 5.2** (Release 13.06.2026) ein zweiter chinesischer Frontier-Coding-Kandidat. Siehe [[../concepts/llm/glm-5.2-zai-coding-model.md]] für Details.
ingest(youtube): ichbinfabian-glm-5.2
GLM 5.2 Deep-Dive (IAmFabian, 14.06.2026, 13:15 Min, 212 views,
16 Likes). Deutscher Reviewer testet live: Flappy Bird (1:20, 481
Zeilen HTML, laeuft) und Newton-Pendel mit echter Physik (9:41, davon
8:27 nur Thinking).
Kernkonzepte:
- Chinas Frontier-Coding-Modelle 2026: Kimi K2.7 + GLM 5.2, beide
Frontier-Niveau bei 10x niedrigeren Preisen
- GLM 5.2: 1M Kontext, MIT-Lizenz (Open Weights 'naechste Woche'),
laut AI Code King Benchmarks nur 6% unter Opus 4.8
- Z.ai Coding Plan (18 USD/Monat Light) als exklusive Distribution
- Light-Plan: 2-3x Kontingent-Mehrverbrauch in Stosszeiten
- OpenRouter-Eintrag 'in den naechsten Tagen' erwartet
- Coding Helper (npx @z_ai/coding-helper) bindet GLM 5.2 in Claude
Code, OpenCode, Cline, Kilo Code, Crush, Factory Droid ein
KRITISCHER IMPLIKAT FUER OPENCLAW:
- Z.ai Coding Plan listet 'Clawdbot/OpenClaw' naamentlich als
unterstuetztes Environment - direkte Integration bereits vorbereitet
- Drei Setup-Optionen:
1. Z.ai Coding Plan sofort (18 USD/Monat, Light)
2. Self-Hosting nach Open Weights (MIT, HuggingFace)
3. Warten auf OpenRouter (Tage)
Aenderungen:
- raw/youtube/2026-06-15_ichbinfabian-glm-5.2-coding-modell.md (neu)
- wiki/concepts/glm-5.2-zai-coding-model.md (neu)
- wiki/architecture/model-routing.md (neue Sektion GLM 5.2 +
Routing-Position + 3 Setup-Optionen)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung: 12 Z.ai/HuggingFace/NPM-Links, 10 Tool-Links
(OpenCode, Claude Code, Cline, Kilo Code, Crush, Droid),
8 Wiki-Cross-References, 4 Konkurrenz-Links (Kimi, OpenRouter,
Open LLM Leaderboard, Anthropic).
2026-06-15 09:50:23 +02:00
ingest(github): openclaw releases v2026.6.6-v2026.6.8
Raw dump of top-5 OpenClaw releases (16.06.2026) via GitHub REST API.
Highlights of the digest:
- v2026.6.8 stable: GLM-5.2 + Claude Haiku 4.5 catalog rows, OpenRouter/Vertex
provider-prefix normalization, Telegram rich-message delivery, QMD memory
search stable in transient mode, raw Memory Wiki source pages as source
evidence, Skill Workshop symlink-write validation
- v2026.6.7-beta.1: Kimi K2.7 Code added, Discord thread-title timeout,
Anthropic thinking replay repair
- v2026.6.6 stable: major security release — exec approvals fail closed on
timeout, MCP stdio hardening, Telegram account-scoped topics
Wiki changes:
- tools/openclaw.md (new): platform reference with version timeline, 10
plattformrelevante changes vs our setup, rollout plan (awaiting Kai approval
for v2026.6.8 update)
- architecture/model-routing.md (updated): GLM-5.2-Support + provider-prefix
normalization for v2026.6.8, distribution matrix updated
- architecture/memory-system.md (updated): QMD stabilization, SQLite-NFS-WAL
detection, raw-Memory-Wiki-source-pages as source evidence, full-reindex
rollback recovery
- concepts/llm/glm-5.2-zai-coding-model.md (updated): OpenClaw v2026.6.8
native provider adapter breaks Z.ai Coding Plan monopoly
- index.md (updated): new tools entry, header datum 12. update, updated
descriptions for memory/routing/glm-5.2, new raw sources row
- log.md (updated): new ingest entry following schema
2026-06-16 20:17:21 +02:00
**Update 2026-06-16:** Mit OpenClaw v2026.6.8 ist GLM 5.2 **nativ in OpenClaw
als Provider-Adapter verfügbar** — vorher war der Zugang auf den Z.ai Coding
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
Plan beschränkt (siehe [[../concepts/llm/glm-5.2-zai-coding-model.md]]). Damit
ingest(github): openclaw releases v2026.6.6-v2026.6.8
Raw dump of top-5 OpenClaw releases (16.06.2026) via GitHub REST API.
Highlights of the digest:
- v2026.6.8 stable: GLM-5.2 + Claude Haiku 4.5 catalog rows, OpenRouter/Vertex
provider-prefix normalization, Telegram rich-message delivery, QMD memory
search stable in transient mode, raw Memory Wiki source pages as source
evidence, Skill Workshop symlink-write validation
- v2026.6.7-beta.1: Kimi K2.7 Code added, Discord thread-title timeout,
Anthropic thinking replay repair
- v2026.6.6 stable: major security release — exec approvals fail closed on
timeout, MCP stdio hardening, Telegram account-scoped topics
Wiki changes:
- tools/openclaw.md (new): platform reference with version timeline, 10
plattformrelevante changes vs our setup, rollout plan (awaiting Kai approval
for v2026.6.8 update)
- architecture/model-routing.md (updated): GLM-5.2-Support + provider-prefix
normalization for v2026.6.8, distribution matrix updated
- architecture/memory-system.md (updated): QMD stabilization, SQLite-NFS-WAL
detection, raw-Memory-Wiki-source-pages as source evidence, full-reindex
rollback recovery
- concepts/llm/glm-5.2-zai-coding-model.md (updated): OpenClaw v2026.6.8
native provider adapter breaks Z.ai Coding Plan monopoly
- index.md (updated): new tools entry, header datum 12. update, updated
descriptions for memory/routing/glm-5.2, new raw sources row
- log.md (updated): new ingest entry following schema
2026-06-16 20:17:21 +02:00
wird die "Warten auf OpenRouter-Listing"-Option (siehe unten) gegenstandslos.
ingest(youtube): ichbinfabian-glm-5.2
GLM 5.2 Deep-Dive (IAmFabian, 14.06.2026, 13:15 Min, 212 views,
16 Likes). Deutscher Reviewer testet live: Flappy Bird (1:20, 481
Zeilen HTML, laeuft) und Newton-Pendel mit echter Physik (9:41, davon
8:27 nur Thinking).
Kernkonzepte:
- Chinas Frontier-Coding-Modelle 2026: Kimi K2.7 + GLM 5.2, beide
Frontier-Niveau bei 10x niedrigeren Preisen
- GLM 5.2: 1M Kontext, MIT-Lizenz (Open Weights 'naechste Woche'),
laut AI Code King Benchmarks nur 6% unter Opus 4.8
- Z.ai Coding Plan (18 USD/Monat Light) als exklusive Distribution
- Light-Plan: 2-3x Kontingent-Mehrverbrauch in Stosszeiten
- OpenRouter-Eintrag 'in den naechsten Tagen' erwartet
- Coding Helper (npx @z_ai/coding-helper) bindet GLM 5.2 in Claude
Code, OpenCode, Cline, Kilo Code, Crush, Factory Droid ein
KRITISCHER IMPLIKAT FUER OPENCLAW:
- Z.ai Coding Plan listet 'Clawdbot/OpenClaw' naamentlich als
unterstuetztes Environment - direkte Integration bereits vorbereitet
- Drei Setup-Optionen:
1. Z.ai Coding Plan sofort (18 USD/Monat, Light)
2. Self-Hosting nach Open Weights (MIT, HuggingFace)
3. Warten auf OpenRouter (Tage)
Aenderungen:
- raw/youtube/2026-06-15_ichbinfabian-glm-5.2-coding-modell.md (neu)
- wiki/concepts/glm-5.2-zai-coding-model.md (neu)
- wiki/architecture/model-routing.md (neue Sektion GLM 5.2 +
Routing-Position + 3 Setup-Optionen)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung: 12 Z.ai/HuggingFace/NPM-Links, 10 Tool-Links
(OpenCode, Claude Code, Cline, Kilo Code, Crush, Droid),
8 Wiki-Cross-References, 4 Konkurrenz-Links (Kimi, OpenRouter,
Open LLM Leaderboard, Anthropic).
2026-06-15 09:50:23 +02:00
**Empfohlene Routing-Position:**
```
OpenClaw Primary Routing
ingest(github): openclaw releases v2026.6.6-v2026.6.8
Raw dump of top-5 OpenClaw releases (16.06.2026) via GitHub REST API.
Highlights of the digest:
- v2026.6.8 stable: GLM-5.2 + Claude Haiku 4.5 catalog rows, OpenRouter/Vertex
provider-prefix normalization, Telegram rich-message delivery, QMD memory
search stable in transient mode, raw Memory Wiki source pages as source
evidence, Skill Workshop symlink-write validation
- v2026.6.7-beta.1: Kimi K2.7 Code added, Discord thread-title timeout,
Anthropic thinking replay repair
- v2026.6.6 stable: major security release — exec approvals fail closed on
timeout, MCP stdio hardening, Telegram account-scoped topics
Wiki changes:
- tools/openclaw.md (new): platform reference with version timeline, 10
plattformrelevante changes vs our setup, rollout plan (awaiting Kai approval
for v2026.6.8 update)
- architecture/model-routing.md (updated): GLM-5.2-Support + provider-prefix
normalization for v2026.6.8, distribution matrix updated
- architecture/memory-system.md (updated): QMD stabilization, SQLite-NFS-WAL
detection, raw-Memory-Wiki-source-pages as source evidence, full-reindex
rollback recovery
- concepts/llm/glm-5.2-zai-coding-model.md (updated): OpenClaw v2026.6.8
native provider adapter breaks Z.ai Coding Plan monopoly
- index.md (updated): new tools entry, header datum 12. update, updated
descriptions for memory/routing/glm-5.2, new raw sources row
- log.md (updated): new ingest entry following schema
2026-06-16 20:17:21 +02:00
├─ Creative / Long-Context (1M) → GLM 5.2 (via OpenClaw provider-adapter)
├─ Speed / Innovation Coding → Kimi K2.7 Code (added in v2026.6.7-beta.1)
├─ Standard Tasks → openrouter/auto → ollama/minimax-m3
ingest(youtube): ichbinfabian-glm-5.2
GLM 5.2 Deep-Dive (IAmFabian, 14.06.2026, 13:15 Min, 212 views,
16 Likes). Deutscher Reviewer testet live: Flappy Bird (1:20, 481
Zeilen HTML, laeuft) und Newton-Pendel mit echter Physik (9:41, davon
8:27 nur Thinking).
Kernkonzepte:
- Chinas Frontier-Coding-Modelle 2026: Kimi K2.7 + GLM 5.2, beide
Frontier-Niveau bei 10x niedrigeren Preisen
- GLM 5.2: 1M Kontext, MIT-Lizenz (Open Weights 'naechste Woche'),
laut AI Code King Benchmarks nur 6% unter Opus 4.8
- Z.ai Coding Plan (18 USD/Monat Light) als exklusive Distribution
- Light-Plan: 2-3x Kontingent-Mehrverbrauch in Stosszeiten
- OpenRouter-Eintrag 'in den naechsten Tagen' erwartet
- Coding Helper (npx @z_ai/coding-helper) bindet GLM 5.2 in Claude
Code, OpenCode, Cline, Kilo Code, Crush, Factory Droid ein
KRITISCHER IMPLIKAT FUER OPENCLAW:
- Z.ai Coding Plan listet 'Clawdbot/OpenClaw' naamentlich als
unterstuetztes Environment - direkte Integration bereits vorbereitet
- Drei Setup-Optionen:
1. Z.ai Coding Plan sofort (18 USD/Monat, Light)
2. Self-Hosting nach Open Weights (MIT, HuggingFace)
3. Warten auf OpenRouter (Tage)
Aenderungen:
- raw/youtube/2026-06-15_ichbinfabian-glm-5.2-coding-modell.md (neu)
- wiki/concepts/glm-5.2-zai-coding-model.md (neu)
- wiki/architecture/model-routing.md (neue Sektion GLM 5.2 +
Routing-Position + 3 Setup-Optionen)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung: 12 Z.ai/HuggingFace/NPM-Links, 10 Tool-Links
(OpenCode, Claude Code, Cline, Kilo Code, Crush, Droid),
8 Wiki-Cross-References, 4 Konkurrenz-Links (Kimi, OpenRouter,
Open LLM Leaderboard, Anthropic).
2026-06-15 09:50:23 +02:00
└─ Frontier Quality → Claude Opus 4.8 / Fable
```
ingest(github): openclaw releases v2026.6.6-v2026.6.8
Raw dump of top-5 OpenClaw releases (16.06.2026) via GitHub REST API.
Highlights of the digest:
- v2026.6.8 stable: GLM-5.2 + Claude Haiku 4.5 catalog rows, OpenRouter/Vertex
provider-prefix normalization, Telegram rich-message delivery, QMD memory
search stable in transient mode, raw Memory Wiki source pages as source
evidence, Skill Workshop symlink-write validation
- v2026.6.7-beta.1: Kimi K2.7 Code added, Discord thread-title timeout,
Anthropic thinking replay repair
- v2026.6.6 stable: major security release — exec approvals fail closed on
timeout, MCP stdio hardening, Telegram account-scoped topics
Wiki changes:
- tools/openclaw.md (new): platform reference with version timeline, 10
plattformrelevante changes vs our setup, rollout plan (awaiting Kai approval
for v2026.6.8 update)
- architecture/model-routing.md (updated): GLM-5.2-Support + provider-prefix
normalization for v2026.6.8, distribution matrix updated
- architecture/memory-system.md (updated): QMD stabilization, SQLite-NFS-WAL
detection, raw-Memory-Wiki-source-pages as source evidence, full-reindex
rollback recovery
- concepts/llm/glm-5.2-zai-coding-model.md (updated): OpenClaw v2026.6.8
native provider adapter breaks Z.ai Coding Plan monopoly
- index.md (updated): new tools entry, header datum 12. update, updated
descriptions for memory/routing/glm-5.2, new raw sources row
- log.md (updated): new ingest entry following schema
2026-06-16 20:17:21 +02:00
**Drei Setup-Optionen** (Stand 16.06.2026 — Update nach OpenClaw v2026.6.8):
1. **OpenClaw Provider-Adapter** (NEU seit v2026.6.8) — nativ in OpenClaw als
Provider routeable, keine externe Subscription nötig. **Empfohlener Pfad.**
2. **Z.ai Coding Plan** (parallel verfügbar, ab 18 USD/Monat) — via
`npx @z_ai/coding-helper` . **Achtung:** Light-Plan hat in Stoßzeiten 2-3×
Kontingent-Verbrauch und lange Wartezeiten.
3. **Self-Hosting** (nach Open Weights Release nächste Woche) — MIT-Lizenz,
Self-Host via [HuggingFace ](https://huggingface.co/zai-org/GLM-5 ).
4. ~~**Warten** auf OpenRouter-Listing~~ — **obsolet** , OpenClaw hat's
nativ, und OpenRouter-Listing erfolgt unabhängig.
ingest(youtube): ichbinfabian-glm-5.2
GLM 5.2 Deep-Dive (IAmFabian, 14.06.2026, 13:15 Min, 212 views,
16 Likes). Deutscher Reviewer testet live: Flappy Bird (1:20, 481
Zeilen HTML, laeuft) und Newton-Pendel mit echter Physik (9:41, davon
8:27 nur Thinking).
Kernkonzepte:
- Chinas Frontier-Coding-Modelle 2026: Kimi K2.7 + GLM 5.2, beide
Frontier-Niveau bei 10x niedrigeren Preisen
- GLM 5.2: 1M Kontext, MIT-Lizenz (Open Weights 'naechste Woche'),
laut AI Code King Benchmarks nur 6% unter Opus 4.8
- Z.ai Coding Plan (18 USD/Monat Light) als exklusive Distribution
- Light-Plan: 2-3x Kontingent-Mehrverbrauch in Stosszeiten
- OpenRouter-Eintrag 'in den naechsten Tagen' erwartet
- Coding Helper (npx @z_ai/coding-helper) bindet GLM 5.2 in Claude
Code, OpenCode, Cline, Kilo Code, Crush, Factory Droid ein
KRITISCHER IMPLIKAT FUER OPENCLAW:
- Z.ai Coding Plan listet 'Clawdbot/OpenClaw' naamentlich als
unterstuetztes Environment - direkte Integration bereits vorbereitet
- Drei Setup-Optionen:
1. Z.ai Coding Plan sofort (18 USD/Monat, Light)
2. Self-Hosting nach Open Weights (MIT, HuggingFace)
3. Warten auf OpenRouter (Tage)
Aenderungen:
- raw/youtube/2026-06-15_ichbinfabian-glm-5.2-coding-modell.md (neu)
- wiki/concepts/glm-5.2-zai-coding-model.md (neu)
- wiki/architecture/model-routing.md (neue Sektion GLM 5.2 +
Routing-Position + 3 Setup-Optionen)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung: 12 Z.ai/HuggingFace/NPM-Links, 10 Tool-Links
(OpenCode, Claude Code, Cline, Kilo Code, Crush, Droid),
8 Wiki-Cross-References, 4 Konkurrenz-Links (Kimi, OpenRouter,
Open LLM Leaderboard, Anthropic).
2026-06-15 09:50:23 +02:00
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
**Ensemble-Kandidat:** GLM 5.2 + Kimi K2.7 als heterogenes Coding-Panel in [[../concepts/llm/llm-model-fusion-ensembles.md]] (unterschiedliche Architektur: Dense vs MoE, unterschiedliche Stärken).
ingest(youtube): ichbinfabian-glm-5.2
GLM 5.2 Deep-Dive (IAmFabian, 14.06.2026, 13:15 Min, 212 views,
16 Likes). Deutscher Reviewer testet live: Flappy Bird (1:20, 481
Zeilen HTML, laeuft) und Newton-Pendel mit echter Physik (9:41, davon
8:27 nur Thinking).
Kernkonzepte:
- Chinas Frontier-Coding-Modelle 2026: Kimi K2.7 + GLM 5.2, beide
Frontier-Niveau bei 10x niedrigeren Preisen
- GLM 5.2: 1M Kontext, MIT-Lizenz (Open Weights 'naechste Woche'),
laut AI Code King Benchmarks nur 6% unter Opus 4.8
- Z.ai Coding Plan (18 USD/Monat Light) als exklusive Distribution
- Light-Plan: 2-3x Kontingent-Mehrverbrauch in Stosszeiten
- OpenRouter-Eintrag 'in den naechsten Tagen' erwartet
- Coding Helper (npx @z_ai/coding-helper) bindet GLM 5.2 in Claude
Code, OpenCode, Cline, Kilo Code, Crush, Factory Droid ein
KRITISCHER IMPLIKAT FUER OPENCLAW:
- Z.ai Coding Plan listet 'Clawdbot/OpenClaw' naamentlich als
unterstuetztes Environment - direkte Integration bereits vorbereitet
- Drei Setup-Optionen:
1. Z.ai Coding Plan sofort (18 USD/Monat, Light)
2. Self-Hosting nach Open Weights (MIT, HuggingFace)
3. Warten auf OpenRouter (Tage)
Aenderungen:
- raw/youtube/2026-06-15_ichbinfabian-glm-5.2-coding-modell.md (neu)
- wiki/concepts/glm-5.2-zai-coding-model.md (neu)
- wiki/architecture/model-routing.md (neue Sektion GLM 5.2 +
Routing-Position + 3 Setup-Optionen)
- wiki/index.md (neuer Concepts-Eintrag + Raw Sources-Eintrag)
- wiki/log.md (Eintrag)
Maximale Verlinkung: 12 Z.ai/HuggingFace/NPM-Links, 10 Tool-Links
(OpenCode, Claude Code, Cline, Kilo Code, Crush, Droid),
8 Wiki-Cross-References, 4 Konkurrenz-Links (Kimi, OpenRouter,
Open LLM Leaderboard, Anthropic).
2026-06-15 09:50:23 +02:00
2026-06-16 11:39:58 +02:00
**Besonderheit:** [Z.ai Coding Plan ](https://z.ai/subscribe ) listet `Clawdbot/OpenClaw` namentlich als unterstütztes Environment — direkte Integration bereits vorbereitet.
## Post-Transformer-Architekturen als künftige Routing-Optionen
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
Während aktuelles Routing primär **innerhalb** des AR-Transformer-Stacks arbeitet (mit Modell-Heterogenität als Variations-Achse), verschiebt DeepMind die Variations-Achse **architektonisch** : hybride Attention+Recurrence (Griffin, Recurrent Gemma, Titans), Diffusions-LLMs, JEPA-Weltmodelle. Details: [[../concepts/llm/post-transformer-llm-architectures.md]].
2026-06-16 11:39:58 +02:00
**Aktuell noch nicht relevant für OpenClaw-Routing** (keine offenen Weights + stabile Provider-Listings), aber mittelfristig zu beobachten:
- **State-Space-Hybride** (Titans, Mamba) als effizientere lokale Embedding-Backends für SSR/Speicher-Layer
- **Diffusions-LLMs** für parallele Code-Completion in nicht-Präfix-Positionen (relevant für Hermes-Agent-Code-Edit-Tasks)
- **JEPA-Weltmodelle** für Subconscious-Agent-„Simulations-Prior" (statt nur Text-Pretraining-Prior)
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
**Status:** Beobachten, nicht adoptieren. Sobald OpenRouter oder ein anderer Provider-Adapter stabile Listings anbietet, evaluieren wie bei [[../concepts/llm/glm-5.2-zai-coding-model.md]].
2026-06-16 13:43:48 +02:00
## Hardware-Frontier als Routing-Constraint
feat: people-pages for all TODO institutions markers
- Created 27 new people-pages for all TODO-marked persons
- Replaced all TODO: people-page markers with wiki-links in institutions
- Cross-referenced back to institutions from each people-page
- 0 TODO markers remaining
New people-pages:
shayne-coplan, jim-zemlin, wang-changhu, yang-bingyang, demis-hassabis,
clement-delangue, thomas-kurian, fei-fei-li, christopher-manning,
sam-altman, ilya-sutskever, moritz-kaminski, elon-musk, nicolas-burtey,
elizabeth-stark, olaoluwa-osuntokun, jan-leike, oren-etzioni,
ali-farhadi, jaime-sevilla, max-tegmark, daniela-rus, jared-kaplan,
alex-atallah, andrew-moore, tuomas-sandholm, mitchell-baker
2026-06-25 21:44:25 +02:00
Mainzers Energie-Argument (20W Gehirn vs. Megawatt-LLM-Cluster, siehe [[../concepts/hardware/neuromorphic-chips-und-quantencomputer.md]]) validiert [[../concepts/llm/ai-value-migration-orchestration.md]] (Aravind: „Token Value per Watt per User"). Implikation für OpenClaw-Routing:
2026-06-16 13:43:48 +02:00
- **Edge-Tasks** (Smart-Home, Mobile, Embedded): langfristig nur mit neuromorphen/photonischen Backends wirtschaftlich — nicht mit Cloud-LLMs
- **Mittelfristig beobachten:** Intel Loihi, IBM TrueNorth, Lightmatter, Luminous Computing — Status 2026?
- **Kurzfristig:** Optimierung bestehender AR-Transformer-Pfade (Caching, Quantisierung, Speculative Decoding) bleibt wichtig, ist aber Endpunkt-Frickelei — Hardware-Diversität ist der eigentliche Spielfeldwechsel.
2026-07-07 10:31:37 +02:00
## Update 2026-07-07: Matthew Berman Model Routing Patterns
**Source:** [[../../raw/youtube/2026-07-07_berman-model-routing.md]]
Matthew Berman's video on model routing provides practical cost-saving patterns that complement our existing architecture. Core finding: **planning vs. execution separation** — use expensive frontier models (Fable) for architecture/spec design, then route code execution to cheap models (GPT 5.5, GLM 5.2, Composer 2.5).
### Key additions to our routing knowledge:
| Pattern | Savings | Relevance to OpenClaw |
|---------|---------|----------------------|
| Manual Copy-Paste | ~60% | Ad-hoc testing, proof-of-concept |
| Cross-Model-Calling | ~70% | Relevant for agent orchestration (subconscious -> execution) |
| Cursor Auto Mode | ~75% | IDE-integrated routing |
| Not Diamond | ~80% | Dedicated routing layer (complementary to our fallback-chain) |
**Coinbase routing on GLM 5.2** validates enterprise adoption of open-weight routing — GLM 5.2 as execution layer behind frontier planning models.
**Cost math:** Fable-only baseline $9.50 -> routed $6.48 (68% savings). Potential >90% with aggressive routing.
See [[../tools/model-routing.md]] for the full tools-level page on routing cost-saving patterns.