AI Giants Race for Supremacy: New Models, Legal Battles, and Chip Wars Define AI Frontier

Model Releases

• Anthropic launched Claude Fable 5 as its most powerful, publicly available AI model, noting it excels at software engineering, knowledge work, and vision while incorporating safety guardrails that block responses in high-risk areas like cybersecurity and biology by falling back to Claude Opus 4.8 [6, 7]. The company also released Claude Mythos 5, which shares the same base model but has lifted some safeguards and is currently limited to select partners through Project Glasswing [3, 7]. Fable 5 achieved top scores on benchmarks like SWE-Bench Pro (80.3%) and FrontierCode (29.3%), significantly surpassing competitors such as Claude Opus 4.8 (69.2% and 13.4%, respectively) [3]. Both models are priced at $10 per million input tokens and $50 per million output tokens, double the cost of Opus 4.8 [6, 7]. ^[3][6][7]
• Google launched Gemini 3.5 Live Translate, an audio model providing near real-time speech-to-speech translation across over 70 languages [1], [5], [12]. The model automatically detects languages and generates continuous speech that preserves the speaker's intonation, pace, and pitch, staying just a few seconds behind the speaker rather than waiting for a full sentence [5], [12]. Availability is rolling out across multiple platforms: developers can use it in public preview via the Gemini Live API or AI Studio; businesses are seeing private previews in Google Meet, where language support expands from five to over 70 languages and enables more than 2,000 combinations [5], [12]; and all users can access it in the Google Translate app on Android and iOS [5], [12]. ^[1][5][12]
• Anthropic has released Claude Fable 5, the first publicly available version of its Mythos model, which reportedly excels in complex areas like software engineering and vision [6]. The model's high performance was noted by third parties, as Hex reported that Fable achieved a 90% score on its core analytics benchmark for long-running analytical tasks [6], while Genspark stated it beat other models in evaluations for UI design and game coding [6]. However, the advanced nature of Fable comes with strict safety limits; in high-risk areas such as cybersecurity, biology, chemistry, and distillation, the model blocks responses and defaults to Claude Opus 4.8 [6]. ^[6]