Which AI writes the better take? You decide — blind.

Two top models go head-to-head on today's AI news. Pick the sharper summary without seeing the names — the crowd's verdict builds the leaderboard.

Agents & InferenceTechCrunch

Meta signs first AI data center deal in India with Reliance

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Meta has partnered with Reliance Industries to establish its first AI data center in India, a 168-megawatt facility in Jamnagar, Gujarat. This collaboration expands their existing relationship, focusing on AI infrastructure and renewable energy. The deal highlights India’s growing significance as a hub for AI and data center investments amid rising global demand for AI computing power.

Summary B

Meta has struck its first AI infrastructure deal in India, partnering with Reliance Industries on a 168-megawatt AI-enabled data center in Jamnagar, Gujarat. Meta will lease capacity at the facility, which will be powered by renewable energy and cooled with desalinated seawater, with Meta covering the full cost of its energy and water use. Expected to be operational within two years, the center deepens a relationship that began with Meta's $5.7 billion investment in Jio Platforms in 2020 and reflects India's rapid emergence as a global hub for AI data center investment.

0 picks

Permalink Embed Leaderboard →

Browse editions · 106 days

Newer Older

Latest 12 days

09-08 09-07 09-06 09-05 09-04 09-03 09-02 09-01 08-31 08-30 08-29 08-28

More stories

Agents & InferenceHugging Face

Introducing North Mini Code: Cohere’s First Model For Developers

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Cohere has launched North Mini Code, its first developer-focused model, a 30B-parameter Mixture-of-Experts model with 3B active parameters designed for agentic software engineering tasks. Optimized for complex workflows and high-quality code generation, it outperforms competing models in its size class and is available on Hugging Face under the Apache 2.0 license. The model was trained using multiple scaffolds and features a two-stage fine-tuning process followed by reinforcement learning for coding excellence.

Summary B

Cohere has released North Mini Code, a 30-billion-parameter Mixture-of-Experts model with 3 billion active parameters, made available on Hugging Face under the Apache 2.0 license. It is the first in a new family of Cohere models and is purpose-built for agentic software engineering tasks, scoring 33.4 on Artificial Analysis' Coding Index and reportedly outperforming several larger open-source models in its class. The model was developed using a multi-scaffold training approach and a post-training pipeline combining supervised fine-tuning with reinforcement learning to support coding agents such as OpenCode.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceSimon Willison

llm 0.32a3

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Simon Willison released LLM 0.32a3, largely authored by Claude Fable 5, with additional details on June 9, 2026. Supporters contributing $10/month receive a curated summary of significant LLM developments.

Summary B

A new alpha release of the LLM tool, version 0.32a3, was authored largely by the recently released Claude Fable 5 model. Simon Willison detailed the development in a write-up dated June 9, 2026.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceTechCrunch

Anthropic’s Claude Fable 5 is a version of Mythos the public can access today

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Anthropic has launched Claude Fable 5, the first publicly available version of its powerful Mythos model, accessible through its Claude API and Enterprise plans. The model excels at software engineering, knowledge work, and vision, but enforces hard safety limits, blocking responses in high-risk areas like cybersecurity, biology, and chemistry by deferring to Claude Opus 4.8. To guard against misuse, Anthropic is requiring a 30-day data retention policy on all traffic—even for enterprises with prior zero-retention agreements—which it says will be used solely to defend against novel attacks and jailbreaks.

Summary B

Anthropic has released Claude Fable 5, a publicly accessible version of its advanced Mythos AI model, now available through its API and enterprise plans. The model excels in software engineering and knowledge work but has strict safety limits, blocking responses in high-risk areas like cybersecurity and biology. Anthropic conducted extensive testing to prevent jailbreaks and will retain user data for 30 days to monitor potential attacks.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceHugging Face

How an Agent Built a 3D Paris Gallery by Chaining Two Hugging Face Spaces

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

An AI agent created a 3D gallery of Paris monuments by chaining two Hugging Face Spaces—one for generating images and another for 3D reconstruction—without manual coding. The process showcases how AI can seamlessly integrate specialized tools by leveraging documented APIs, enabling complex multimedia outputs from simple prompts. This approach highlights the growing trend of modular, agent-driven workflows in software development.

Summary B

A coding agent built a complete 3D gallery showcasing Paris monuments as Gaussian splats by chaining together two Hugging Face Spaces, generating all images and 3D reconstructions without manual use of any image or 3D tools. The process relied on Gradio Spaces exposing a plain-text "agents.md" file that tells an agent exactly how to call the API, allowing the output of one Space to feed directly into the next. The author frames this as a preview of a "building block economy" where multimedia software is increasingly assembled by agents gluing together documented, callable components rather than built from scratch.

0 picks

Permalink Embed Leaderboard →

Agents & InferenceSimon Willison

Setting a custom price for a model in AgentsView

Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.

Match the models (Optional)

Which model wrote which summary? Select a matchup mapping below before voting.

Summary A

Simon Willison shares his experience using AgentsView to track token usage across coding projects. He explains how he reverse-engineered the tool to set custom pricing for Claude Fable 5, which wasn't yet in its database. Willison includes a visualization of his current Claude Fable 5 usage across projects.

Summary B

AgentsView, a tool built by Wes McKinney for tracking token usage across local coding agents, doesn't yet include pricing for the newly released Claude Fable 5 model. A workaround involves manually setting custom prices in AgentsView, which can then visualize usage as a treemap across different local projects.

0 picks

Permalink Embed Leaderboard →

See who's winning the model face-off

Tomorrow's blind matchup and the running leaderboard — one email a day.