Frontier and open model releases, benchmarks, and what changed between versions. 51 stories so far. Latest, on : Dream Labs Raises $90M Seed, Opens Dream-1 Robot API at $3 an Hour.
Dream Labs launched with a $90M seed led by B Capital and Kindred Ventures and a public API for Dream-1, a robot foundation model priced at $3/hour that controls unseen two-armed robots zero-shot.
Meta's Muse Wonder costs $4 in / $20 out per million tokens with 1M context. Wonder Full scores 79.0 on DeepSWE and 57.5 on SWE-Marathon in Meta's scorecard but trails Claude Opus 5.5 on six rows.
Nous Research confirmed a $90M Series B at a $1.5B valuation led by Robot Ventures, per TechCrunch, bringing total funding to $158M. Hermes Agent has 24M+ clones by Nous’s estimate.
Google DeepMind’s Gemini 4 Argon opens via Fairwind with a 1M output token limit, intro pricing at $2/$10 per 1M tokens, and cyber-defense benchmarks including DeepSWE 77.9% and CWE-bench 68%.
TypeSafe AI opened Jev early access: typed decisions at $0.042/MTok input, output free. Official blog claims 40-200x faster. The model cannot generate text.
With the xAI Grok 4.6 release, Elon Musk's AI startup takes on GPT-5 and Claude 4 with massive context windows and real-time social data access.
The Meta Muse Glimmer release delivers a 30B local AI model that outperforms Google's Gemma4-31B, running under 20GB VRAM on consumer GPUs.
Meta Muse Spark 1.2 lands with major upgrades to agentic AI workflows, matching GPT-5.5 benchmarks while maintaining aggressive developer pricing.
MiniMax H3 open source release delivers 2K video with native stereo audio, but keeps its critical context-refinement module behind a proprietary API.
OpenAI launches Health in ChatGPT, integrating personal medical data with HIPAA-compliant guardrails to challenge Apple and Google in consumer health AI.
OpenAI Health in ChatGPT launched to integrate AI into consumer healthcare. Learn how OpenAI navigates medical accuracy, regulation, and HIPAA privacy.
OpenAI Presence represents the shift to ambient AI agents and spatial computing. Here is what OpenAI's new framework means for the industry.
Moonshot AI's Kimi K3 agentic benchmark performance ranks second only to Fable 5, proving Chinese labs are closing the gap in autonomous AI execution.
Google Gemini Flash models get a massive upgrade with Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber, targeting speed, cost, and security.
The iOS 27 public beta is now live, giving users early access to Apple's revamped Siri AI. See how its on-device agentic power compares to Google.
Chinese AI lab Moonshot releases Attention Residuals, a drop-in Transformer upgrade claiming 1.25x compute savings, plus 48B MoE model weights under MIT license.
Anthropic's Claude 4 family—Opus and Sonnet—redefined AI coding and agentic workflows. Here's how it stacks up against GPT-5 and Gemini 2.5.
Anthropic launched Claude Opus 4 and Sonnet 4 with dominant coding benchmarks. Here's how they stack up against GPT and Gemini in real-world use.
Anthropic's Claude 4 family—Opus and Sonnet—delivers dominant coding and agentic capabilities that force OpenAI and Google to respond. Full analysis inside.
Anthropic shipped seven Claude models in eight months, dominating coding benchmarks and undercutting rivals on price. Here's how Claude stacks up against GPT-5 and Gemini 3.
Anthropic launches Claude Opus 4 and Sonnet 4 with dominant coding benchmarks and agentic capabilities. Here's how they stack up against GPT-5 and Gemini 2.5.
Anthropic's Claude 4 model family has delivered massive capability jumps in coding, reasoning, and agentic tasks. Here's what it means for the AI race.
Anthropic releases Claude Opus 4 and Claude Sonnet 4 with hybrid reasoning, agentic coding dominance, and extended thinking — reshaping the AI model landscape.
Anthropic's Claude Opus 4 and Sonnet 4 bring hybrid reasoning, elite coding benchmarks, and agentic workflows that raise the bar in the frontier model race.
Anthropic launches Claude Opus 4 and Sonnet 4 with elite agentic coding, extended thinking, and 7-hour autonomous task endurance. Here's what the benchmarks say.
Anthropic launches Claude Opus 4 and Sonnet 4 with breakthrough coding, reasoning, and agentic capabilities. Here's what matters and how it stacks up.
Anthropic's Claude 4 family — from Opus to Haiku — is rewriting the rules on coding, agents, and reasoning. Here's how it stacks up against GPT-5 and Gemini.
Anthropic's Claude 4 Opus and Sonnet models bring elite coding, 7-hour agentic sessions, and extended thinking. Here's how they stack up against GPT-5 and Gemini 2.5.
Anthropic's Claude Opus 4 and Sonnet 4 deliver breakthrough agentic coding, hybrid reasoning, and extended thinking — reshaping the frontier model landscape.
From Claude 4 to enterprise plugins and a $380B valuation, Anthropic's relentless 2025 model cadence is forcing OpenAI and Google into reactive mode.
Anthropic's Claude 4 family — Opus and Sonnet — set new benchmarks for agentic coding and extended thinking, fundamentally reshaping the AI race against OpenAI and Google.
Anthropic's Claude Opus 4 and Sonnet 4 dominate coding benchmarks against GPT-4.1 and Gemini 2.5 Pro. We break down the numbers and what they mean for devs.
Google launched a wave of AI products including Gemini 2.5 Flash-Lite, Gemma 3n, Gemini CLI, and Imagen 4 — the biggest coordinated AI release of 2025 so far.
Google's Gemini 3.1 Pro and Anthropic's Claude Sonnet 4.6 both dropped this week. Here's what actually matters and who's winning the AI model war.
Gemini 3.1 Pro, Claude Sonnet 4.6, GPT-5.2, and a wave of open-source challengers. Here's every major AI model dropping right now.
From Anthropic's $30B raise to Gemini 3.1 Pro vs GPT-5.2 model wars, the AI industry is moving faster than ever. Here's what's shaping the landscape.
OpenAI, Anthropic, and Google are locked in the most intense AI model competition yet. Here's what mid-2025's launch wave means for the industry.
Gemini 3.1 Pro, Claude Opus 4.6, and GPT-5.2 are reshaping the AI competitive landscape. Here's who's winning, who's scrambling, and what it means for developers.
Anthropic and Google both shipped major AI models this week. Here's what Claude Sonnet 4.6 and Gemini 3.1 Pro actually mean for developers and users.
Google's Gemini 2.5 Pro and Anthropic's Claude dominate the 2025 AI race. Here's who's winning, who's losing, and what it means for developers.
Google's Gemini 3.1 Pro and Anthropic's Claude Sonnet 4.6 just dropped within days of each other. Here's what it means and who's actually winning.
HuggingFace Transformers v5 release includes 1,200 commits and major breaking API changes. Here's the migration path every AI team needs to follow.
OpenAI's GPT-5.2 reportedly cites Grokipedia as a source, raising questions about AI training data practices and how competing systems influence each other.
SGLang inference optimization project spins out as RadixArk with $400M valuation and Accel backing. What it means for the AI inference market.
OpenAI's Higgsfield case study confirms GPT-5 and Sora 2 are powering video creation tools. Here's what this quiet reveal means for AI developers.
OpenAI's partner case study reveals GPT-5 and Sora 2 are powering Higgsfield's cinematic video creation platform. Here's what this means for AI video tools.
Higgsfield AI video startup reaches $1.3B unicorn valuation on $200M ARR, a rare revenue-backed milestone in a space where most competitors are still burning cash.
vLLM achieves 2,200 tokens per second per H200 GPU serving DeepSeek using wide expert parallelism. A technical breakthrough for MoE model inference costs.
Google Veo 3.1 introduces vertical video generation and improved consistency controls. Here's what the update means for creators and the AI video race.
Lawrence Berkeley Lab's new LLM AI copilot manages 230,000+ variables at the Advanced Light Source accelerator. Here's why this matters for scientific infrastructure.
NVIDIA RTX PRO 5000 72GB Blackwell GPU is now available, giving AI developers enough VRAM to run 70B+ parameter models locally. Here's what it means for agentic AI.