Tag: AI models
Claude Opus 5.5: lower prices, faster output and gains on agentic benchmarks
Anthropic launches Claude Opus 5.5 at $4/$20 per million tokens, with faster output, a fast mode and gains on Terminal-Bench and GDPval-AA.
Google I/O 2026: Gemini 3.5 Flash, Antigravity 2.0 and managed agents in the Gemini API
At I/O 2026, Google launched Gemini 3.5 Flash, Antigravity 2.0 with a CLI and SDK, Managed Agents in the Gemini API, and proposed WebMCP as a web standard.
Chrome 148: the Prompt API brings an on-device language model to the web, and CSS levels up
Chrome 148 hit stable with the Prompt API (on-device model with constrained output), name-only container queries, revert-rule and Web Serial on Android.
GPT-5.5 hits the API: 1M-token context, doubled pricing and High cyber risk
GPT-5.5 brings a 1.05M-token context and an xhigh reasoning effort, costs $5/$30 per million tokens and is rated High in bio and cyber.
MacBook Pro with M5 Pro and M5 Max: Apple targets people running LLMs locally
Up to 128GB of unified memory, 614GB/s and a Neural Accelerator in every GPU core: what the new MacBook Pro changes for running models locally.
Claude Opus 4.6: 1M context, effort levels and agent teams in Claude Code
Anthropic releases Claude Opus 4.6 with a 1M-token context in beta, 128k output, adaptive thinking, effort levels and agent teams in Claude Code.
NVIDIA Rubin: six chips, an NVL72 rack and what it means for developers
NVIDIA unveils the Rubin platform, with the Rubin GPU, Vera CPU and KV cache storage, and promises inference up to 10x cheaper than Blackwell.
Gemini 3 Flash: Google's fast model becomes the default and targets coding work
Gemini 3 Flash costs $0.50/$3 per million tokens, becomes the default in the Gemini app and AI Mode, and enters preview in the API, Gemini CLI and Antigravity.