xAI Launches Grok Code Fast 1 at $0.20 Per Million Tokens—70.8% on SWE-Bench with 256K Context
Elon Musk’s xAI just entered the coding assistant war with a model priced 87% below GPT-5—and every major IDE integrated it before the…
Alibaba’s Qwen3.8-Max Launches with 2.4 Trillion Parameters—But Zero Official Benchmarks
Alibaba just dropped a 2.4-trillion-parameter model claiming it’s “second only to Claude Fable 5.” Five days later,…
Runway Launches Dev API on July 17—Single Integration Point for Gen-4.5, Seedance, GPT Image 2, and ElevenLabs
Runway just made its competitors’ models available through its own billing dashboard. One integration now gives you Gen-4.5, Seedance,…
Mistral Releases Leanstral 1.5—119B Open-Source Model Hits 100% on miniF2F, Solves 587 of 672 Putnam Problems
Mistral’s new theorem prover found 5 unreported bugs across 57 repositories by proving code mathematically wrong—including an integer…
Anthropic Launches Claude Sonnet 5 at $2/$10 Per Million Tokens—Default Model for All Free and Pro Users Hits 63.2% on Agentic Coding Benchmark
The AI model that can autonomously navigate browsers, execute terminal commands, and complete multi-step workflows is now free for every…
OpenAI’s o3 Scores 87.5% on ARC-AGI, 96.7% on AIME, and 2727 Codeforces Elo—Announced December 20 with ‘Deliberative Alignment’ Safety Framework
OpenAI just tripled its previous model’s score on a benchmark specifically designed to be unsolvable by current AI—and they did it using…
GitLab Survey: 78% of Developers Code Faster with AI—But Software Delivery Hasn’t Accelerated
78% of developers write code faster with AI tools. Yet 79% say their organization’s software delivery speed hasn’t improved. The…
Anthropic Launches Claude Tag on June 23—Slack AI Agent Writes 65% of Product Team’s Code, Runs on Claude Opus 4.8
Anthropic just revealed that 65% of its own product team’s code comes from an AI agent—and they’ve opened that same system to…
Microsoft Shifts GitHub Copilot to Token-Based Billing—Forrester Calls It ‘The New Gateway Drug’ as Azure Foundry Adds 11,000+ Models
Microsoft just turned AI coding assistance into a metered utility—and the enterprises that budgeted for unlimited copilots are about to…
Figure AI’s Helix 02 Replaces 109,504 Lines of C++ with 10M-Parameter Neural Network—Humanoid Robot Completes 61-Action Dishwasher Task in 4 Minutes
Figure AI just mass-deleted over 100,000 lines of carefully engineered robotics code. A 10-million-parameter neural network now does the job better.