Anthropic ships Claude Opus 4.8 with parallel subagent workflows and 2.5x fast mode

Anthropic has released Claude Opus 4.8, making it available in Claude Code and on third-party APIs such as AI/ML API, according to official and third-party sources. The company flagged the launch as a major upgrade for the flagship model.

The new release promises significantly better coding and agentic capabilities. According to TestingCatalog, internal tests show roughly 4× fewer code flaws going unnoticed compared with Opus 4.7. The model also introduces a Fast Mode running at 2.5× speed that is now 3× cheaper, while keeping the same $5/$25-per-million-token pricing.

Alongside the model, Anthropic rolled out a research-preview feature called dynamic workflows. As described by the Claude Devs account, the system can write an orchestration script on the fly and spin up a large fleet of coordinated subagents in parallel to tackle complex tasks. Users can trigger the capability by including the word "workflow" in a prompt.

Community observers reacted enthusiastically. Min Choi noted that Opus 4.8 passes tricky reasoning evaluations such as the "car wash & strawberry test," and highlighted that "Mythos-class models are coming in weeks." Others had anticipated the drop, with Choi earlier speculating about a possible same-day launch against an OpenAI model.

The release bundles a frontier model upgrade with infrastructure for multi-agent execution, signaling Anthropic’s push to make Claude a central engine for autonomous software development.