Live telemetry · daily brief
GAWK
by nativerse
● ARCHIVE
2026-09-06
4 changes in the LMArena top 3
What verifiably moved in the AI ecosystem in the last 24h. Every number traces to a public source.
What moved
- ■New #1 on LMArena: claude-fable-5.1-max overtook claude-opus-5-max.
SDK adoption5 notable shifts across the six registries
Package install volume is where developers place real bets — distinct from where labs are shipping marketing. pypistats.org, crates.io, hub.docker.com · view all →
+365.3k
Docker Hub: ollama/ollama
+365.3k all-time pulls day-over-day
+13.3k
VS Code Marketplace: saoudrizwan.claude-dev
+13.3k cumulative installs day-over-day
Agent frameworks3 agent frameworks moved >10% in the last week
Agent frameworks rise and fall on weekly download velocity and open-issue trajectory, not GitHub stars. pypistats.org, pypistats.org, pypistats.org · view all →
-81%
CrewAI · -81%
570k weekly downloads · 58k stars
-40%
OpenAI Agents · -40%
4.2M weekly downloads · 29k stars
-24%
Pydantic AI · -24%
1.2M weekly downloads · 20k stars
Tracks the `pydantic-ai` meta-package; the lean `pydantic-ai-slim` variant is roughly 4× larger but users typically compare the meta number.
Model Usagetencent/hy4-preview holds #1 · biggest mover: deepseek/deepseek-v4-flash-vision-exp +11 to #38 · biggest drop: nvidia/nemotron-3-super-120b-a12b -13
OpenRouter rankings reflect API-first developer spend; direct customers like consumer ChatGPT are invisible by construction. openrouter.ai · view all →
+11
deepseek/deepseek-v4-flash-vision-exp
was #49 on 2026-08-30
OpenRouter request volume reflects developer API spending, not end-user adoption. Biased toward API-first workflows that route through OpenRouter — direct OpenAI / Anthropic / Google customers who never use OpenRouter are invisible.
-13
nvidia/nemotron-3-super-120b-a12b
was #30 on 2026-08-30
—
#1 tencent/hy4-preview
#1 tencent/hy4-preview
—
#2 z-ai/glm-5.3-flash
#2 z-ai/glm-5.3-flash
—
#3 deepseek/deepseek-v4-flash-0731
#3 deepseek/deepseek-v4-flash-0731
Tool Health
All tools operational, no incidents in the last 24h (vs 0 yesterday)
Why this matters · Provider outages and degradations cause retry storms upstream. Tracking the 7-day shape catches flapping providers before they page you.
claude-code
Operational
openai-api
Operational
codex
Operational
copilot
Operational
windsurf
Operational
cursor
Operational
Benchmark movers
4 changes in the LMArena top 3
Why this matters · Public benchmarks are gameable, but rank shuffles still hint at where the frontier is genuinely moving.
New to top 3: claude-fable-5.1-max
#1 · anthropic · 1514 Elo
claude-opus-5-high
+1 Elo (now 1505) · anthropic
claude-opus-5-max
down 2 to #3 · anthropic
Dropped from top 3: claude-opus-4-6-high
was #3 · anthropic
Top HN stories
Top 5 on HN in the last 24h
Why this matters · What developers debate on Hacker News often previews which models, frameworks, and patterns they'll actually adopt next.
LLMs as a Cognitive Virus
140 points · 125 comments
GPT-6 Astra on robot arms
86 points · 44 comments
Finite time blowup for an averaged three-dimensional Navier-Stokes equation (2014)
53 points · 21 comments
OKF Agent Memory – Git-native persistent memory for AI coding agents
40 points · 14 comments
Show HN: We Beat MLPerf: Modern Storage for KV Offload and LLM Training
34 points · 1 comments
Notable AI Lab activity
9 labs moved in the last 24h
Why this matters · GitHub event volume on a lab's own repos is the cleanest publicly-verifiable proxy for engineering activity we have.
LangChain
+6 events (now 188) · San Francisco, US
OpenAI
+26 events (now 144) · San Francisco, US
Hugging Face
+26 events (now 122) · New York, US
Ollama
−6 events (now 92) · San Francisco, US
Microsoft Research
−30 events (now 42) · Redmond, US
DeepSeek
−61 events (now 22) · Hangzhou, CN
Weights & Biases
−46 events (now 9) · San Francisco, US
New mover: BAAI (Beijing Academy of AI)
7 events · Beijing, CN
Dropped off: INRIA
was 16 events · Paris, FR
Get this board in your inbox — one email, every morning (UTC).
Subscribe free