Claude Opus 4.5 Launches: Safer SOTA Coding and Agents, Now Cheaper and More Efficient

Anthropic released Claude Opus 4.5, a safer, more capable, and more efficient flagship model for coding, agents, and complex computer use, available across apps, API, and major clouds at reduced pricing. The model posts state-of-the-art results on key benchmarks, offers controllable effort levels for speed/cost versus depth, and shows better token efficiency and long-horizon performance. Upgrades to the developer platform and apps (Claude Code, Chrome, Excel, long-chat) aim to translate these gains into everyday productivity.

Key Points

Claude Opus 4.5 launches with state-of-the-art coding, agentic, and computer-use capabilities, broad availability, and lower pricing ($5/$25 per million tokens).
Token efficiency and control improve via a new effort parameter, context compaction, memory, and advanced tool use—matching or exceeding Sonnet 4.5 with dramatically fewer tokens.
Benchmark results and internal tests show significant gains: SOTA on SWE-bench Verified, leadership on multiple agentic and coding benchmarks, and superior performance on an internal engineering exam.
Safety is emphasized: Opus 4.5 is presented as Anthropic’s most robustly aligned model, with industry-leading resistance to strong prompt injection attacks.
Product ecosystem upgrades (Claude Code Plan Mode and desktop app, long-chat persistence, Chrome and Excel integrations) leverage Opus 4.5’s long-horizon and computer-use strengths.

Sentiment

The Hacker News sentiment is largely mixed but leans towards cautious optimism. There's considerable excitement and positive reception for the significant price reduction and removal of usage caps, which many see as game-changing for usability and production viability. However, a strong undercurrent of skepticism and cynicism exists, primarily driven by past experiences with model degradation ('nerfing'), doubts about benchmark integrity, and ongoing concerns about transparency and privacy. Many users are eager to test the model's new capabilities, especially in coding, while others remain wary or have already switched to competitors.

In Agreement

The 3x price drop to $5/$25 per MTok makes Opus 4.5 viable for production workloads and daily use, significantly improving its competitive position.
The removal of Opus-specific usage caps and increased overall limits address a major pain point for users, making Opus 4.5 accessible for consistent daily work.
Claims of state-of-the-art prompt injection resistance are significant for deploying agents with tool access, marking a practical safety advancement.
The model's strong performance in coding, agentic workflows, and specific tasks like orchestrating visual pipelines or analyzing legacy codebases is highly valued by developers.
Opus 4.5's token efficiency means that despite a higher per-token cost than some models, its ability to achieve results with fewer tokens can lead to lower overall costs.
Many users appreciate Anthropic's focus on developers and the improvements to Claude Code's Plan Mode and tool use, enhancing multi-model workflows.
The competitive market, spurred by this release, is seen as beneficial for consumers.

Opposed

Many users express strong skepticism about model performance consistency, citing a common pattern of initial strong performance followed by degradation ('nerfing') of models like Sonnet 4.5, potentially due to throttling or model changes.
There's significant distrust in benchmarks, with users suggesting they are 'gamed' or don't reflect real-world performance, and pointing out misleading data visualization in the announcement.
Some users speculate that the price drop and speed increase might indicate Opus 4.5 is a smaller, more quantized, or sparse (MoE) model compared to Opus 4.1, potentially compromising untargeted performance.
Concerns about Anthropic's data privacy practices, particularly the use of user prompts for training, persist among some former and current users.
Despite the price drop, some users still find Opus 4.5 significantly more expensive than competitors like Gemini 3 Pro and Grok 4, with some perceiving Gemini as performing better for general tasks or even certain coding scenarios.
A segment of the community critiques Anthropic's ethical stance, particularly regarding lobbying efforts that could lead to regulatory capture or restrict open-weight models.
Some users feel that agentic workflows, even with improved models, are not yet suitable for production-level code at scale and require too much manual oversight.
Questions arise about the role and effectiveness of Haiku, which seems to occupy a 'weird middle point' without clear advantages for many tasks compared to smaller or larger models.