Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Vals AI

AI overview

Sign in and the AI will write an overview from our coverage.

Headlines · 3

  1. Anthropic's Claude Haiku 5.5 debuts as fastest model tested; Sonnet 5.5 tops coding index

    Anthropic released Claude Haiku 5.5 on October 7, replacing the year-old Haiku 4.5; Artificial Analysis scored it 43 on its Intelligence Index (13th among available models, ahead of all "fast" models) and measured 237 tokens per second, the fastest it has tested, though average time to first token is 337 seconds. API pricing is $0.10/$0.01/$0.50 per million input/cached/output tokens, with all prices 5x above 100k tokens. Claude Sonnet 5.5, launched September 28 at Sonnet 5's per-token prices ($2/$0.10/$10 per million), ranked first on Artificial Analysis' Coding Agent Index (68), second on its Intelligence Index (56) and second on Vals AI's Vals Index (67.04%), but Artificial Analysis says another model matches or beats it at every effort level at lower cost per task, and at max effort Sonnet generates about 193,000 output tokens per task. Anthropic says Sonnet 5.5 is the first Sonnet to block reasoning extraction without falling back to another model, part of an effort to make distillation harder.

    The Batch · DeepLearning.AI · 🔥 7
  2. Z.ai releases GLM-Legal, claims top open-weight legal model

    Z.ai released GLM-Legal, a legal-domain model built on the lightweight GLM-5.3-Flash base, saying it matches frontier closed models on mainstream legal benchmarks for workflows such as contract review, evidence analysis and compliance planning. According to Z.ai and Vals AI, GLM-Legal scored a 19.2% full-pass rate on Harvey LAB, a 12.5 percentage-point gain over its base model, and ranks as the top open-weight model on Harvey's Legal Agent Benchmark and LegalBench. These performance and ranking claims come from the developer and related parties and have not been independently reproduced.

    Z.ai（智谱） · 🔥 0
  3. Study: AI agent teams cost 1.8–5.1x more for little quality gain

    Evals company Vals AI tested GPT-6 Sol and Claude Opus 5.5 on its Vibe Code Bench both as single agents and as agent teams, at medium and maximum reasoning effort. Teams cost between 1.8x and 5.1x more than single agents, yet only one of four comparisons showed a statistically significant improvement: GPT-6 Sol at medium reasoning, where the team scored 7.3 points higher; at maximum reasoning, neither Sol's nor Opus 5.5's team gained a real advantage. The report also notes that in two of Anthropic's own tests with Opus 5.5, quality gains shrank as more agents were added, while OpenAI researcher Noam Brown has said multi-agent systems mainly buy speed rather than better quality.

    The Decoder · 🔥 0

Experience and discussion from the community

Share my Vals AI experienceAsk about Vals AI

Nobody has shared their experience with Vals AI yet.