Agent framework 1.0 cluster: LangGraph + LlamaIndex + CrewAI + AutoGen + OpenAI Agents SDK + Anthropic Agent SDK all hit milestones in one week

8 sources

The week of 2026-08-06 saw a cluster of agent framework milestones: LangGraph 0.5 + LangGraph Studio GA (LangChain), Workflows GA + LlamaDeploy 1.0 (LlamaIndex), Flows GA + Crews 1.0 (CrewAI), AutoGen 0.4 actor-model refactor (Microsoft), OpenAI Agents SDK 1.0, Anthropic Agent SDK 1.0. The Decoder frames the cluster as 'every agent framework ships a 1.0 / GA milestone in the same week'. The pattern: agent orchestration is now a buyer-side market with four open-source frameworks (the open-source quad) and two vendor SDKs (the vendor pair). Each milestone represents a stable, production-ready surface.

Why now

The cluster is the editorial framing that turns six independent framework milestones into a single coherent 'agent orchestration is now a stable buyer-side market' story for creators — useful because the comparison is most informative when shown side-by-side.

Why it is worth publishing

Demo potential: a side-by-side of the same multi-agent workflow (a recurring weekly brief) implemented across all six frameworks, measuring time-to-first-output, observability surface, and lock-in.

Evidence basis

The Decoder + IT之家 weekly roundups + six independent framework primary sources

“Six agent framework milestones shipped in one week — LangGraph, LlamaIndex, CrewAI, AutoGen, OpenAI Agents SDK, Anthropic Agent SDK — and the buyer-side market for agent orchestration just stabilized across four open-source frameworks and two vendor SDKs.”

Angle

Use the cluster to introduce the 'agent orchestration is now a stable buyer-side market' pattern — four open-source frameworks + two vendor SDKs — and use that lens to compare approaches side-by-side.

Format

Long-form explainer

Demo idea

Record a 16-minute comparison explainer: 2 min intro on 'stable buyer-side market' framing, then 2 min per framework (LangGraph / LlamaIndex / CrewAI / AutoGen / OpenAI Agents SDK / Anthropic Agent SDK), then a 4-min side-by-side of the same multi-agent workflow across all six.

Platform notes

Each framework's release notes frame its release against the competitive set the maintainer cares about; The Decoder and IT之家 are editorial framing layers, not independent verification. Confirm any specific primitive or API surface against the underlying framework docs before stating it on the record.

Usable claims

LangChain shipped LangGraph 0.5 framework (graph-based agent orchestration with persistent state, human-in-the-loop interrupt/resume, time-travel debugging) + LangGraph Studio GA.
LlamaIndex shipped Workflows GA (event-driven agent orchestration with step-level async / sync / streaming handlers, automatic instrumentation hooks) + LlamaDeploy 1.0.
CrewAI shipped Flows GA (event-driven orchestration with persistent state across crews) + Crews 1.0 (multi-agent crews with role / goal / backstory primitives).
Microsoft shipped AutoGen 0.4 — actor-model architecture refactor (replacing the group-chat manager abstraction), async-first design, cross-language support (Python + .NET).
OpenAI shipped Agents SDK 1.0 — handoffs between specialized agents, guardrails (input/output validation hooks), built-in tracing dashboard, OpenAI Traces UI.
Anthropic shipped Agent SDK 1.0 — first-class Skills integration, computer use API GA (mouse / keyboard control of virtualized desktop), workbench CLI for local debugging.

Evidence pipeline

From the news

Breakdown

Six agent framework milestones shipped in one week — the editorial framing ('agent orchestration is now a stable buyer-side market') is useful but turns into a vendor directory if you don't introduce the underlying pattern. This explainer uses the cluster to introduce the 'four open-source frameworks + two vendor SDKs' pattern and uses that lens to compare approaches side-by-side, while keeping the per-framework primitives honest (LangGraph on graph state, LlamaIndex on event-driven steps, CrewAI on role-based multi-agent, AutoGen on actor-message passing, OpenAI on handoff + guardrails, Anthropic on Skills + computer use).

Sources

Risks

Use The Decoder and IT之家 as media-type corroboration, but read the underlying framework docs for any specific primitive or API surface before stating it on the record. Verify specific capability claim against the underlying vendor docs and the actual license / pricing matrix before stating it on the record; do not paraphrase per-platform pricing or license terms into specific dollar figures or commercial-use clauses.
Docs confirm the framework features exist but specific primitive signatures and protocol details beyond the captured summary are not in this pass. Verify specific capability claim against the underlying vendor docs and the actual license / pricing matrix before stating it on the record; do not paraphrase per-platform pricing or license terms into specific dollar figures or commercial-use clauses.

Demo ideas

Side-by-side multi-agent workflow: same recurring weekly brief across all six frameworks, plot time-to-first-output, observability surface, lock-in.
Decision tree: 'which framework for which use case' (graph-based / human-in-the-loop → LangGraph, RAG-heavy / document-grounded → LlamaIndex, role-based multi-agent → CrewAI, cross-language → AutoGen, OpenAI model native → OpenAI Agents SDK, Anthropic model native → Anthropic Agent SDK).
Open-source quad vs vendor pair comparison: plot each framework on openness vs vendor-coupling.