AI model intelligence feed
Frontier model releases, benchmarks, and price changes — in one live feed.
Built for AI engineers, founders, and analysts who need to know what changed before everyone else. Tracks the mainstream models and agents — releases, benchmarks, leaderboards, capital activity, and provider status across the frontier labs and the open-source race.
Recent releases
View all →Claude 4.8 Opus
released 2026-05-28
Anthropic's flagship hybrid-reasoning model — Opus 4.8 pushes coding and AI-agent workflows further (agentic coding 69.2% vs 64.3% on 4.7; multidisciplinary reasoning 57.9% vs 54.7%). 1M-token context, $5/$25 per MTok pricing held flat from Opus 4.7. Introduces Dynamic Workflows for Claude Code (research preview). GA on Anthropic API (`claude-opus-4-8`), AWS Bedrock, Google Vertex, Microsoft Foundry, and GitHub Copilot.
- Context
- 1,000,000
- License
- proprietary
Gemini 3.5
released 2026-05-20
Google DeepMind's next-gen Gemini — positioned as "frontier intelligence with action". Built for complex agentic workflows. Announced at Google I/O 2026.
- Context
- 2,000,000
- License
- proprietary
Qwen3.7-Max
released 2026-05-20
Alibaba's flagship agent model — 1M-token context, extended-thinking mode, 56.6 on the Artificial Analysis Intelligence Index v4.0 (5th overall, #1 Chinese). 50.8% on Terminal-Bench Hard. Designed for long-horizon agent workloads (hundreds-to-thousands of steps). Closed-weight, $2.50/$7.50 per 1M tokens.
- Context
- 1,000,000
- License
- proprietary
Gemini Omni
released 2026-05-20
Multimodal Gemini variant introduced at Google I/O 2026 — unified text, image, audio, and video processing in a single model.
- Context
- 1,000,000
- License
- proprietary
GPT-5.5
released 2026-04-23
OpenAI's smartest and most intuitive model — successor to GPT-5, with major coding, research, and document workflow gains. GPT-5.5 Pro variant also available for heavier reasoning.
- Context
- 400,000
- License
- proprietary
DeepSeek-V4-Flash
released 2026-04-22
Smaller, faster sibling to DeepSeek-V4-Pro. Same 1M context window with a much lighter 284B / 13B-active MoE.
- Context
- 1,000,000
- Params
- 284B (13B active)
- License
- MIT
- Source
- open
AI agents
All agents →The deployed-product layer atop raw models — coding agents, browser agents, autonomous assistants.
Replit Agent
Replitcoding
Replit's in-browser coding agent. Agent 4 (Mar 2026) introduced parallel task forking that auto-resolves merge conflicts ~90% of the time.
Hermes Agent
Nous Researchgeneral·open source
Open-source AI agent from Nous Research with a built-in learning loop — creates skills from experience, persists knowledge, builds a model of its user across sessions. ~60K stars in two months.
Cursor
Anyspherecoding
AI-native code editor, fork of VS Code. Hit $2B ARR in February 2026. Composer mode for natural-language multi-file refactors.
Codex
OpenAIcoding
OpenAI's coding agent — runs in CLI, IDE, and cloud-hosted runtimes. Drives the Plus/Pro coding workflow.
Claude Code
Anthropiccoding
Anthropic's CLI coding agent. Top of the SWE-bench Verified leaderboard at 80.8% on Opus 4.6. Lives in the terminal — reads files, executes commands, navigates repos, applies multi-file edits autonomously.
OpenClaw
Erik Steinbergergeneral·open source
Open-source autonomous AI agent (formerly Clawdbot, Moltbot — lobster theme). Self-hosted; supports Claude, GPT, KIMI, MiMo, Qwen 3, Llama 4, Mistral, plus local models via Ollama. ~347K GitHub stars.
Latest news
All news →Codex is becoming a productivity tool for everyone
Codex is being positioned as a productivity tool for a broad range of knowledge workers, with the Next Era of Knowledge Work report highlighting its use for AI-powered research, data analysis, workflow automation, and content creation. It matters because the report frames Codex as moving beyond a niche coding assistant into a general-purpose work tool that can speed up multiple white-collar tasks.
- capital
Alphabet plans to raise $80 billion to pay for AI buildout
Alphabet plans to raise $80 billion by selling stock to fund its AI buildout. The size of the planned capital raise underscores how expensive large-scale AI infrastructure has become for major tech companies.
- release
OpenAI SDK v2.40.0
2.40.0 (2026-06-01) Full Changelog: v2.39.0...v2.40.0 Features api: Add Amazon Bedrock Responses support Bug Fixes api: allow setting bedrock api keys on the client directly (4d5bfde)
- infra
Nvidia chases $200B CPU market with AI agent PCs from Microsoft, Dell, and HP
Nvidia is aiming at the $200 billion CPU market by pushing AI agent PCs with Microsoft, Dell, and HP, betting that it has found a way to bring AI agents to consumers and enterprises easily, safely, and usefully. If successful, it would expand Nvidia beyond GPUs into mainstream personal computers and put it in direct competition with established CPU vendors.
- release
OpenAI SDK v2.39.0
v2.39.0, released 2026-06-01, adds workload identity entries to audit logs, an additional_tools item in responses, and makes ActionSearch.query optional. These changes improve observability and response structure while fixing a schema issue that could have made ActionSearch integrations unnecessarily strict.
- capital
Anthropic Files Confidentially For IPO
Anthropic said Monday it confidentially filed for a proposed IPO, without disclosing the offering size or listing venue, shortly after a $65 billion Series H round that lifted its post-money valuation to $965 billion. The filing intensifies the race among major AI companies to reach public markets, and if Anthropic follows a SpaceX-like timeline, it could debut as soon as August after regulators review the confidential filing and the company submits its public prospectus.
- capital
Anthropic has officially filed to go public
Anthropic has filed a draft registration statement confidentially with the U.S. Securities and Exchange Commission, kicking off the IPO process after months of speculation about whether it or OpenAI would go public first. The filing comes after Anthropic’s latest fundraise valued it at $965 billion post-money, ahead of OpenAI’s $852 billion valuation, underscoring how large the eventual public offering could be.
How we used Gemini to build Google I/O 2026
Google says Googlers used Gemini to help build Google I/O 2026. The notable detail is that the company is showcasing its own AI tools being used in the production process, though the excerpt provides no specific numbers or model variants.
Introducing Mellum2: A 12B Mixture-of-Experts Model by JetBrains
JetBrains introduced Mellum2, a 12B mixture-of-experts model. It matters because the MoE design suggests JetBrains is targeting stronger performance efficiency for code-focused AI while keeping the parameter count relatively compact.
- capital
AI is blowing up music. How should the Grammys handle it?
Harvey Mason Jr., CEO of the Recording Academy, said AI is now “omnipresent” in music production as Deezer says more than 50,000 AI-generated songs are uploaded every day and tools like Suno are becoming mainstream in musicians’ workflows. The Grammys still exclude AI music from eligibility, making the Recording Academy’s rules a live test case for how major music institutions handle generative AI as it becomes harder to detect and filter.