code agents

$anthropic’s-claude-haiku-4.5-matches-may’s-frontier-model-at-fraction-of-cost$

Anthropic’s Claude Haiku 4.5 matches May’s frontier model at fraction of cost

AI, AI alignment, AI benchmarks, AI coding, AI development tools, AI models, AI programming, AI safety, Amazon Bedrock, Anthropic, API, Biz & IT, Claude Haiku, Claude Sonnet, code agents, GitHub, google cloud, large language models, machine learning, Vertex AI / Mike M. / October 15, 2025

And speaking of cost, Haiku 4.5 is included for subscribers of the Claude web and app plans. Through the API (for developers), the small model is priced at $1 per million input tokens and $5 per million output tokens. That compares to Sonnet 4.5 at $3 per million input and $15 per million output tokens, and Opus 4.1 at $15 per million input and $75 per million output tokens.

The model serves as a cheaper drop-in replacement for two older models, Haiku 3.5 and Sonnet 4. “Users who rely on AI for real-time, low-latency tasks like chat assistants, customer service agents, or pair programming will appreciate Haiku 4.5’s combination of high intelligence and remarkable speed,” Anthropic writes.

Claude 4.5 Haiku answers the classic Ars Technica AI question, “Would the color be called ‘magenta’ if the town of Magenta didn’t exist?”

On SWE-bench Verified, a test that measures performance on coding tasks, Haiku 4.5 scored 73.3 percent compared to Sonnet 4’s similar performance level (72.7 percent). The model also reportedly surpasses Sonnet 4 at certain tasks like using computers, according to Anthropic’s benchmarks. Claude Sonnet 4.5, released in late September, remains Anthropic’s frontier model and what the company calls “the best coding model available.”

Haiku 4.5 also surprisingly edges up close to what OpenAI’s GPT-5 can achieve in this particular set of benchmarks (as seen in the chart above), although since the results are self-reported and potentially cherry-picked to match a model’s strengths, one should always take them with a grain of salt.

Still, making a small, capable coding model may have unexpected advantages for agentic coding setups like Claude Code. Anthropic designed Haiku 4.5 to work alongside Sonnet 4.5 in multi-model workflows. In such a configuration, Anthropic says, Sonnet 4.5 could break down complex problems into multi-step plans, then coordinate multiple Haiku 4.5 instances to complete subtasks in parallel, like spinning off workers to get things done faster.

For more details on the new model, Anthropic released a system card and documentation for developers.

Anthropic’s Claude Haiku 4.5 matches May’s frontier model at fraction of cost Read More »

Hugging Face clones OpenAI’s Deep Research in 24 hours

On Tuesday, Hugging Face researchers released an open source AI research agent called “Open Deep Research,” created by an in-house team as a challenge 24 hours after the launch of OpenAI’s Deep Research feature, which can autonomously browse the web and create research reports. The project seeks to match Deep Research’s performance while making the technology freely available to developers.

“While powerful LLMs are now freely available in open-source, OpenAI didn’t disclose much about the agentic framework underlying Deep Research,” writes Hugging Face on its announcement page. “So we decided to embark on a 24-hour mission to reproduce their results and open-source the needed framework along the way!”

Similar to both OpenAI’s Deep Research and Google’s implementation of its own “Deep Research” using Gemini (first introduced in December—before OpenAI), Hugging Face’s solution adds an “agent” framework to an existing AI model to allow it to perform multi-step tasks, such as collecting information and building the report as it goes along that it presents to the user at the end.

The open source clone is already racking up comparable benchmark results. After only a day’s work, Hugging Face’s Open Deep Research has reached 55.15 percent accuracy on the General AI Assistants (GAIA) benchmark, which tests an AI model’s ability to gather and synthesize information from multiple sources. OpenAI’s Deep Research scored 67.36 percent accuracy on the same benchmark.

As Hugging Face points out in its post, GAIA includes complex multi-step questions such as this one:

Which of the fruits shown in the 2008 painting “Embroidery from Uzbekistan” were served as part of the October 1949 breakfast menu for the ocean liner that was later used as a floating prop for the film “The Last Voyage”? Give the items as a comma-separated list, ordering them in clockwise order based on their arrangement in the painting starting from the 12 o’clock position. Use the plural form of each fruit.

To correctly answer that type of question, the AI agent must seek out multiple disparate sources and assemble them into a coherent answer. Many of the questions in GAIA represent no easy task, even for a human, so they test agentic AI’s mettle quite well.