Sign in
← AI PulseJul 23, 2026

Deep · news · Single-source brief

Anthropic Upgrades Claude Opus to Version 4.6

Anthropic has released Claude Opus 4.6, an upgraded model with enhanced agentic coding, computer use, tool use, search, and financial analysis capabilities.

By Illumora Editorial

Source · Jul 22, 2026, 12:06 AM · On Illumora · Jul 23, 2026, 1:43 AM

Media from the primary source — shown here so you can stay on Illumora.

Rewritten from one allowlisted primary — not independent enterprise reporting. Lanes →

Brief drafted by Illumora’s editorial model from the linked primary source. Ops desk reviews flagged pieces. How we write →

Read the source →Anthropic News (priority) — Claude Opus 4.6
Save

Anthropic has announced the release of Claude Opus 4.6, an upgrade to its flagship model. This new version is designed to improve performance across various domains, including agentic coding, computer use, tool use, search, and finance. It is available today on claude.ai, via the Claude API, and on major cloud platforms.

Key Points

  • Claude Opus 4.6 features a 1M token context window in beta.
  • The model achieves the highest score on the Terminal-Bench 2.0 agentic coding evaluation.
  • It leads all other frontier models on Humanity’s Last Exam, a multidisciplinary reasoning test.
  • On GDPval-AA, an evaluation for economically valuable knowledge work, Opus 4.6 outperforms OpenAI’s GPT-5.2 by approximately 144 Elo points and its predecessor, Claude Opus 4.5, by 190 points.
  • Opus 4.6 performs better than any other model on BrowseComp, which measures online information retrieval.
  • The model's pricing remains $5 per million input tokens and $25 per million output tokens.
  • New features include adaptive thinking and effort controls for developers, along with compaction for longer API tasks.

Context

According to Anthropic, Claude Opus 4.6 improves upon its predecessor's coding skills by planning more carefully, sustaining agentic tasks for longer, and operating more reliably in larger codebases. It also features enhanced code review and debugging capabilities. The model's improved abilities extend to everyday work tasks such as financial analysis, research, and document creation. Within Cowork, Opus 4.6 can multitask autonomously. The model's system card indicates an overall safety profile that is comparable to or better than other frontier models, with low rates of misaligned behavior across safety evaluations.

Why It Matters

For developers and practitioners, Claude Opus 4.6 offers advancements in agentic capabilities and long-context performance, potentially enabling more complex and sustained automated workflows. The introduction of effort controls provides a mechanism to balance intelligence, speed, and cost for specific tasks.

What To Do

  • Test Claude Opus 4.6 via the Claude API using claude-opus-4-6 to explore its enhanced agentic coding and reasoning capabilities.
  • Experiment with the new /effort parameter to adjust the model's thinking depth, balancing performance with cost and latency for different tasks.
  • Review the model's system card for detailed information on its safety profile and evaluation methodologies.
  • Note the 1M token context window in beta for applications requiring extensive context processing.

Keep Exploring

/atlas/claude-family /techniques/multishot /techniques/system-user-separation /studio?pack=foundation