Anthropic debuts Claude Opus 4.8 with agentic workflows
Anthropic has released Claude Opus 4.8 and introduced dynamic workflows for Claude Code, positioning the AI assistant for more autonomous software development as the company continues to expand its product ecosystem. The model is now available through claude.ai, Claude Code, and the Claude API, with standard pricing unchanged at $5 per million input tokens and $25 per million output tokens.

The new model is roughly four times less likely to let coding errors pass unremarked, a shift Anthropic frames around improved “honesty” in self-assessment. Claude Opus 4.8 moved from 64.3 to 69.2 on SWE-bench Pro, but the company’s CTO highlighted the behavioral improvement, stating the advancement he keeps returning to is honesty.
Dynamic Workflows Enable Parallel Agent Execution
Dynamic workflows are launching in research preview for Claude Code users on Enterprise, Team, and Max plans. The feature allows Claude to plan work, run hundreds of parallel subagents in a single session, verify outputs, and report back to the user. This marks a significant shift from single-prompt coding assistance toward autonomous workflow orchestration.
The capabilities have already drawn attention in early use. A recent report documented how Claude Code autonomously created a 70-agent, four-phase research pipeline from a single prompt in ultracode mode, demonstrating the platform’s ability to generate complex workflows without explicit human orchestration.
Claude Code’s dynamic workflows feature is designed for software tasks that stretch beyond a single prompt, file, or short coding exchange, with Claude now able to plan the work, distribute it across hundreds of parallel subagents, and verify outputs before handing work back to the user. However, the same autonomous capabilities have raised concerns about cost controls, with observers flagging the lack of built-in spend caps as a major cost-risk issue for users.
Fast Mode Gets Cheaper and Broader Model Support
Fast mode is now 2.5 times the speed and three times cheaper than for previous models. Additionally, fast mode now supports Claude Opus 4.7, with significantly faster output token generation at premium pricing when users opt in with the appropriate configuration.
Anthropic has launched cache diagnostics in public beta, allowing the API to report a cache_miss_reason explaining where the prompt cache prefix diverged from the previous turn, giving developers more visibility into performance optimization.
Managed Agents Gain Self-Hosted Options and Enhanced Tools
Self-hosted sandboxes are now available for Claude Managed Agents, as an alternative to running tool execution in Anthropic’s infrastructure. Claude adds new Managed Agents capabilities with self-hosted sandboxes and MCP tunnels, giving teams more control over execution and private network access.
The web search tool now returns richer SEC filing data, making it easier to ground financial research agents, earnings analysis, and due-diligence workflows in primary sources with citations. Large outputs from agent_toolset and MCP tools exceeding 100K tokens are now automatically spilled to a file in the sandbox, with the model receiving a truncated preview with the file path.
Platform Expansion and Subscription Changes
Anthropic has launched Claude Platform on AWS, bringing the Claude API to Anthropic-managed infrastructure accessible through AWS, with AWS billing and IAM authentication, including access to the full Messages API, Files API, Message Batches API, Claude Managed Agents, Agent Skills, code execution, and tool use.
Claude releases Microsoft 365 add-ins for Excel, PowerPoint, and Word in general availability, with Outlook now in public beta for paid plans, extending the assistant’s reach across productivity applications.
Subscription structures are also changing. Starting June 15, 2026, programmatic usage of Claude via subscription plans moves to a separate monthly credit pool. Agent SDK and claude -p usage will draw from a new dedicated credit starting June 15, separate from your subscription’s interactive usage limits, though interactive usage through Claude.ai, Claude Code in the terminal, and Claude Cowork remains unaffected.
Legal Tools and Model Updates
Claude releases 20+ new legal MCP connectors and 12 practice-area plugins, expanding how law firms and in-house teams work across research, contracts, discovery, matter management, and legal aid. The company noted that legal professionals have become the most engaged Claude Cowork users of any knowledge-work function since the first legal plugin launched earlier this year.
Claude Opus 4.8 shows improvements over Opus 4.7 in coding, agentic skills, reasoning, and practical knowledge work tasks. Meanwhile, Anthropic announced the deprecation of the Claude Sonnet 4 model and the Claude Opus 4 model, with retirement on the Claude API scheduled for June 15, 2026.
Future Releases and Broader Context
Anthropic is continuing to test Claude Mythos Preview under Project Glasswing with a small number of organizations using it for cybersecurity work, with Mythos-class models expected to come to customers in the coming weeks after stronger cyber safeguards are implemented.
The releases come as Anthropic continues to expand its developer events. Code w/ Claude events are scheduled for Tokyo on June 5-6, with all Day 1 keynotes and breakout sessions to be streamed live, following successful events in London and San Francisco.
Key Facts
- Claude Opus 4.8 released with 4x fewer coding error pass-throughs and improved SWE-bench Pro score (64.3 to 69.2)
- Dynamic workflows in research preview enable hundreds of parallel subagents in a single Claude Code session
- Fast mode now 2.5x faster and 3x cheaper than previous models
- Self-hosted sandboxes now available for Claude Managed Agents
- Starting June 15, 2026, programmatic usage (Agent SDK, claude -p) moves to separate credit pool
- Claude Sonnet 4 and Claude Opus 4 scheduled for retirement June 15, 2026
- Microsoft 365 add-ins (Excel, PowerPoint, Word) now generally available; Outlook in public beta
- 20+ new legal MCP connectors and 12 practice-area plugins released
- Claude Mythos Preview testing continues with expected release in coming weeks
Sources
- Anthropic releases Claude Opus 4.8 for coding agents – EdTech Innovation Hub
- Claude Platform – Claude API Docs
- Claude Code Autonomously Spawns 70 Research Agents – AI Weekly
- Anthropic Splits Claude Subscriptions: What Changes for Indie Hackers on June 15 – DevToolPicks
Sources
- Anthropic releases Claude Opus 4.8 for coding agents | ETIH EdTech News — EdTech Innovation Hub
- Anthropic releases Claude Opus 4.8, promising a more honest model – Techzine Global
- Claude Updates by Anthropic – May 2026 – Releasebot
- Claude Platform – Claude API Docs
- Anthropic Splits Claude Subscriptions: What Changes for Indie Hackers on June 15