Latest
Altman, Amodei, Hassabis and Musk pledge to slow AI development pace Google DeepMind launches Gemini 3.8 Live models for real-time voice conversations OpenAI Foundation pays to mine bankrupt biotech firms for medical data Salesforce launches Koa, a reasoning model built on Nvidia's open Nemotron Startup founded by Anthropic and METR alumni raises $40M to audit AI agents Chinese AI lab Moonshot aims to double revenue to $2 billion by year-end Anthropic researcher resigns, warning of a race toward superintelligence Ex-EPA officials say Trump rollbacks let AI data centers pollute more Anthropic opens its models to outside safety evaluators like METR New Mexico court fines lawyer $5,000 for AI-fabricated witnesses in murder appeal
Telegram

Models

Google DeepMind launches Gemini 3.8 Live models for real-time voice conversations

The two speech-to-speech models handle background tasks and visual context mid-conversation, and Extended Thinking topped Artificial Analysis's Speech to Speech Quality Index with a score of 82.6.

  • Gemini 3.8 Live Extended Thinking scored 82.6 on Artificial Analysis's Speech to Speech Quality Index, ranking first.
  • Gemini 3.8 Live Extended Thinking scored 97.7% on Big Bench Audio and 68.6% on the tau-Voice benchmark.
  • Gemini 3.8 Live automatically detects and switches between 97 supported languages during a conversation.

Latest All stories →

Who is ahead right now All leaderboards →

General questions, writing, reasoning

LMArena

Rank# Model Arena score
1 Claude Fable 5.1 Max Anthropic 1,508
2 Claude Opus 5 Max Anthropic 1,505
3 Claude Opus 5 High Anthropic 1,505
4 Claude Opus 4.6 High Anthropic 1,503
5 Claude Opus 4.6 Anthropic 1,498

Writing and fixing code

LMArena

Rank# Model Arena score
1 GPT 6 Astra Max OpenAI 1,800
2 Claude Fable 5.1 Max Anthropic 1,758
3 Claude Opus 5 Max Anthropic 1,687
4 qwen3.8-max-0902 Alibaba 1,681
5 Kimi K3 Max Moonshot AI 1,674

Cost per million tokens

OpenRouter

Rank# Model Per 1M tokens
1 Mistral Nemo Mistral $0.022
2 Ling 3.0 Flash inclusionAI $0.032
3 Granite 4.0 Micro IBM $0.041
4 DeepSeek V4 Flash Latest DeepSeek $0.055
5 Qwen3.7 Flash Qwen $0.055

Updated 2026-09-16 · sources named on each board

By section

Models Releases, benchmarks, capability jumps

Research Papers that change what is possible

Business Funding, deals, revenue, layoffs

Policy Regulation, safety, copyright, courts

Tools Agents, coding, products that ship

Hardware Chips, data centres, energy