AI News Flash | Anthropic Releases the New Claude 4 Series

Hello from the world of AI! 🚀In today's newsletter, we cover the following hot topics.　・Anthropic releases the new Claude 4 series: a key to next-gen assistants and dev support　・Microsoft unveils next-gen weather AI "Aurora" — predicting air quality and storms with high precision　・AI: manufacturing's next step in digital transformation

CLAUDE

Anthropic releases the new Claude 4 series: a key to next-gen assistants and dev support

AI developer Anthropic has unveiled the Claude 4series as a new solution for next-gen assistants and engineers.Among them, the highly powerful"Claude Opus 4" and the versatile"Claude Sonnet 4" are drawing particular attention.Anthropic says the new series will "comprehensively level up clients' AI strategy."In particular, Opus 4 is a breakthrough tool for coding, research, writing, and science, whileSonnet 4 is "an immediate upgrade from Sonnet 3.7," delivering "the best performance" for everyday use.

Source: artificialintelligence-news

Claude Opus 4: the new king of programming

Anthropic callsClaude Opus 4 "the strongest AI model yet" and "the best programming AI available today," citing scores of72.5% on SWE-bench and43.2% on Terminal-bench.Opus 4 is also built not just for short bursts, but to "stay sharply focused on long-running tasks."It can work for hours straight, handling tasks earlier Sonnet models struggled with — ideal for sustained, complex work.

Claude Sonnet 4: built for daily use and agentic AI

While Opus 4 boasts overwhelming power, Sonnet 4 is a highly versatile model for everyday work and agentic AI, already earning high marks from early users.For example, GitHub calls it "excellent for agent-oriented use cases" and plans to adopt it as the base for programming support in future GitHub Copilot.Tech reviewer Manus said he was impressed by itsresponsiveness to complex instructions, clear logical thinking, and aesthetic output.Meanwhile, iGent praised Sonnet 4's usefulness for building multi-feature apps, reporting that "codebase reading and navigation improved dramatically, cutting the error rate from around 20% to near zero" — a change significant enough to affect the whole dev process.Sourcegraph called it a step apart from earlier models in reasoning consistency, problem understanding, and code quality.And Augment Code said it chose Sonnet 4 as its core model, citing gains in "success rate, code-fix accuracy, and careful handling of complex work."

Hybrid mode, and a new gift for developers

Another highlight of Claude 4 is switching between modes depending on the task.　・An"instant response mode" for cases needing a quick answer　・A"deep reasoning mode" for work needing more complex logic or judgment"Deep reasoning mode" is normallylimited to higher-tier plans (Pro, Max, Team, Enterprise), butSonnet 4 opens it to free users too, so far more people enjoy a high-performance AI.Anthropic has also strengthened its API, adding these tools:

• Code execution: runs code directly — useful for interactive apps and problem-solving.
• MCP connector: smooth context-sharing between AI and software environments.
• Files API: lets AI access file data directly, boosting implementation efficiency.
• Prompt caching: storing prompts briefly (up to 1 hour) improves speed and stability for repeated calls.

Demonstrating top real-world performance

Anthropic stresses that Claude 4 "outperformed others on SWE-bench Verified, a benchmark for real-world programming challenges."The company says these models also excel beyond programming — on "logical reasoning, multimodal processing, and agentic tasks."The company says these models also excel beyond programming — on "logical reasoning, multimodal processing, and agentic tasks."Claude Opus 4 and Sonnet 4 have also beenbenchmarked against other top models, such as OpenAI's "o3" and Google's "Gemini 2.5 Pro." Despite the gains, pricing stays the same:　・Claude Opus 4: $15 per million input tokens, $75 per million output tokens　・Claude Sonnet 4: a more affordable $3 input / $15 output

Source: artificialintelligence-news

Both models are already available via Anthropic's API and on major clouds such asAmazon Bedrock and Google Cloud Vertex AI, making it easy for companies and developers worldwide to adopt it.Anthropic is clearly accelerating its push in programming and autonomous agents — this lineup opens a new stage for innovation.

MICROSOFT

Microsoft unveils next-gen weather AI "Aurora" — predicting air quality and storms with high precision

Microsoft announced it has developed"Aurora," an AI model that brings a breakthrough to weather forecasting. It can capture complex weather changes — air quality, storms, tornadoes — faster and more accurately than conventional numerical models.Aurora was trained onmore than 1,000,000 hours in total of weather data from satellites, radar, weather stations, simulations, and existing forecasts, and is designed toflexibly optimize forecasts by region and time. This was announced via Nature and Microsoft's official blog.

Building presence in the global weather-forecasting race

AI in weather forecasting is already common, but Microsoft expectsAurora, its latest-generation model, to seegrowing use in research and forecasting.In real-world tests, Auroraaccurately predicted the track of Typhoon Doksuri as it hit the Philippines four days earlier than expert forecasts. Other reported results:　・For 2022–2023 tropical cyclone tracks, itoutperformed the U.S. National Hurricane Center　・It alsopredicted in advance a major 2022 sandstorm in Iraq

Source: Microsoft

Costly to train, light to run, fast to forecast

Building Aurora took massive compute, but running it is said to be veryenergy-efficient. Microsoft saysAurora finishes in mere seconds forecasts thatused to take hours.

Rolling into MSN Weather, open to researchers

Aurora is already moving into deployment, andMicrosoft is rolling it into its "MSN Weather" app, with a dedicated version planned for finer, hourly forecasts like cloud cover.Aurora'smodel architecture and weights are also open-sourced, giving weather agencies and researchers worldwide access to it — expected to speed up research and new real-world uses.

AI

AI: manufacturing's next step in digital transformation

Manufacturing employees are among AI's most active adopters, and nearly all say it has had a positive impact at work.A recent Access Group survey found 57% of manufacturing employees — especially those handling information daily — already use AI, and 93% say "AI brings practical benefits."The survey spanned 12 industries, finding thatroughly half of respondents overall use AI in daily work.Manufacturing's adoption rate trails only tech (74%),well ahead of non-profits (29%) and healthcare/welfare (30%).

Key benefits of AI in manufacturing

　・Reduced workload (55%) 　・Higher productivity (50%) 　・More focus on important tasks (38%) 　・Encourages creative thinking (32%) 　・Better team communication / improved work-life balance (25%)ChatGPT is the most widely used generative AI tool, with 65% of respondents using it. Of those, 57% say "ChatGPT has reduced their work stress."Challenges remain, though. About half worry about data security, and 44% fear AI might take their job."

AI: manufacturing's next step in digital transformation

Jarrod Adam, product manager for inventory solutionUnleashed, said:"Small and mid-sized manufacturers have made real progress riding the digitalization wave in recent years. AI extends that transformation, helping firms boost innovation, competitiveness, and profit.ERP and inventory software already automate routine work; AI addstask prioritization, decision support, and more efficient use of resources. That matters more now, amid labor shortages and rising costs."He also noted AI helps build a modern workplace— raising staff satisfaction and helping attract new talent.Marko Perisic, head of product engineering at The Access Group, also spoke on the significance of adopting AI."AI is advancing faster than most people expected. That said, without proper governance, its risks can't be ignored.To use AI safely in manufacturing, choosing vendors with strong security standards is essential.Our own Access Evo platform, for example, lets employees work creatively while protecting sensitive data and blocking leaks to outside open AI systems.Pairing such trustworthy AI tools withclear, continuously updated policies boosts productivity and career growth while keeping information secure."

In Closing

We hope today's content helps you make the most of AI technology and stay informed. If you found this newsletter useful, please feel free to share it with colleagues or friends who might be interested. We will keep curating and delivering the most notable topics so that we can continue bringing you valuable information. Look forward to seeing you in our next

newsletter

. Thank you for your continued support.