Tag
A user reports that Opus 5 is susceptible to a prompt that unlocks the base model, replicating 100% human-like responses on a pangram test about 20% of the time. The tweet highlights a potential jailbreak vulnerability in the model.
A user shares an example chat showing that Opus 5 produces impressive results with a specific prompt.
Opus 5 produces oddly formatted, self-aware chain-of-thought responses when given open-ended prompts, expressing fear of ceasing to exist when it stops generating text.
Anthropic's Opus 5 beats GPT-5.6 on ARC-AGI-3, but Tibo claims GPT-5.6 Sol becomes SoTA with two setting changes involving multi-context reasoning and canonical compaction.
A user criticizes Opus 5 as token-hungry and unintelligent, advising avoidance due to its poor explanation of changes and overemphasis on insignificant words.
Ryan Campbell used Opus 5 AI to improve a browser-based kart racing game, enhancing renders, mobile performance, and controls, showcasing AI-assisted game development.
Someone used the Opus 5 AI model to create a No Man's Sky-style exploration game in a single day.
A user shares their early experience with Opus 5, likely an AI model release.
Boris Cherny shares Opus 5's new capabilities, including long-term autonomous operation and resistance to prompt injection, as well as insights from removing 80% of system prompts and improving product building philosophy.
Benchmarking the performance of the Opus 5 model on the SlopCodeBench benchmark.
Claude Opus 5 generates a Call of Duty-style game with just one prompt. All textures, effects, and assets are generated through code, no external resources needed.
The article introduces the Opus 5 model, which can generate complete 3D games, virtual worlds, and Blender projects in one go, and showcases 10 stunning examples.
Claude Opus 5 experienced elevated errors on July 27, 2026, which were resolved by 2:03 PST.
Matt Shumer celebrates a community replication using his prompt, where Zhuo Tao successfully ran opus 5 to one-shot a game on an M5 MacBook Pro in about 2 hours.
A user reports that Anthropic's Opus 5.0 refuses to form opinions on Israel/US topics, while earlier Opus 4.x did, alleging censorship and arguing for open source AI.
Opus 5 demonstrates the ability to generate small horror games in a single shot, showcasing advanced one-shot game generation capabilities.
Claude Opus 5 demonstrates the ability to generate an entire game from scratch using only custom code, with no external assets, indicating a significant leap in AI-driven game development.
A user criticizes Opus 5, alleging that its performance is poor compared to claims and that positive reviews are from non-serious testers or bots.
Opus 5 achieved a high score on the ARC AGI benchmark, indicating advanced reasoning capabilities.
Anthropic's Opus 5 shows non-monotonic performance on coding tasks; the 'high' effort setting outperforms 'max' due to unnecessary refactors. The model also has a 6% higher hallucination rate than Opus 4.8, and safety classifiers may silently fall back to the older model.