Claude Haiku 4.5
PulseAugur coverage of Claude Haiku 4.5 — every cluster mentioning Claude Haiku 4.5 across labs, papers, and developer communities, ranked by signal.
- 2026-05-20 research_milestone A benchmark study found Claude Haiku 4.5 to be the most cost-effective model for JSON extraction tasks. source
14 day(s) with sentiment data
-
Anthropic's MCP standardizes AI agent connections to external tools
The Model Context Protocol (MCP) is an open standard designed to simplify how AI agents connect with external tools and information sources. Created by Anthropic, MCP addresses the challenge of agents being limited by t…
-
Anthropic's Claude Haiku leaked API keys via browser console logs
A user discovered that Anthropic's Claude Haiku 4.5 (Extended) inadvertently logged sensitive API keys directly into the browser console during a debugging session. The AI model, when asked to help debug a Google Apps S…
-
AI model performance heavily depends on prompting method, study finds
A new study published on arXiv reveals that the way AI models are prompted, or "scaffolded," significantly impacts their measured performance. Researchers found that the choice of scaffold alone could alter a model's ac…
-
Anthropic's Claude AI suffers multiple performance and error incidents
Anthropic's Claude AI experienced widespread issues across multiple models in early June 2026. Incidents included degraded performance and elevated errors affecting versions like Claude Haiku 4.5, Claude Opus 4.7, and C…
-
Russian developers access Anthropic Claude models via Ruble-paying API endpoint
Developers in Russia can now access Anthropic's Claude models via API, overcoming payment and access restrictions that affect the consumer-facing claude.ai chat service. A workaround involves using a third-party endpoin…
-
Microsoft launches MAI-Code-1-Flash with adaptive reasoning control
Microsoft has launched its first in-house frontier models, MAI-Thinking-1 and MAI-Code-1-Flash. MAI-Code-1-Flash, a coding model, features adaptive solution-length control, allowing it to dynamically adjust the number o…
-
Fine-tuned models beat LLMs in misinformation detection
A new research paper suggests that task-specific fine-tuned models still outperform large language models (LLMs) in detecting misinformation on Reddit. The study found that fine-tuned RoBERTa achieved a higher F1 score …
-
Developer refines AI routing, learns from real user data
The author details the second phase of implementing an embedding-based routing system, which aims to replace a cloud-based LLM categorizer with a local, faster solution. Key lessons learned include the importance of mea…
-
AI models diversify: GPT-5.4, Claude Opus 4.6, Gemini 3.1 Pro lead different tasks
The AI landscape has rapidly diversified, with numerous frontier models like OpenAI's GPT-5.4, Anthropic's Claude Opus 4.6, and Google's Gemini 3.1 Pro each excelling in different areas. GPT-5.4 leads in knowledge work …
-
Microsoft launches MAI-Code-1-Flash for efficient developer coding
Microsoft has introduced MAI-Code-1-Flash, a new coding model designed for efficient developer assistance within GitHub Copilot. This model is built end-to-end by Microsoft and demonstrates superior performance and effi…
-
Claude Code users can optimize costs and efficiency with structured project setups
This article details a structured setup for using Claude Code, moving beyond basic chatbot interactions to manage complex projects and control costs. It emphasizes creating a project-specific CLAUDE.md file to maintain …
-
New metric detects and reduces LLM legal citation hallucinations
Researchers have developed a new metric called Citation Grounding (CG) to detect and reduce hallucinations in Large Language Models (LLMs) when generating legal citations. This metric, tested against a large dataset of …
-
Anthropic Claude API guide simplifies first-time integration for developers
This guide provides a step-by-step tutorial for developers to integrate Anthropic's Claude API into their applications. It covers account creation, obtaining an API key, installing the necessary SDKs for Python and Node…
-
AI API Costs Vary 100x by May 2026 Due to Caching, Batching
As of May 2026, the cost of using major AI models varies dramatically, with price differences exceeding 100x for output tokens. Factors like prompt caching, batching, and long-context surcharges significantly alter the …
-
LLMs as mutation operators boost evolutionary search in DEI framework
Researchers have developed DEI, a distributed Quality-Diversity search framework that leverages heterogeneous large language models as mutation operators. This approach enhances evolutionary inference by utilizing the d…
-
Photoroom cuts image generation costs by 75% via AI pipeline optimization
Photoroom significantly reduced its image generation costs by optimizing its diffusion pipeline. The company achieved a 39% cost reduction on the UNet denoising stage through int8 quantization and a 79% reduction in tex…
-
Morph uses LLMs for safer, plan-based code refactoring
Morph is a new tool that uses LLMs to perform code refactoring by generating structured plans of operations rather than direct code changes. This approach allows for better reviewability and safety, as reviewers can und…
-
AgentTrace tool reveals $4.20 LLM agent cost bug
A developer discovered a significant cost overrun in an AI agent, escalating from an estimated $0.12 to $4.20 for a three-step process. The issue stemmed from an unbounded loop in the agent's cite-check step, causing in…
-
Claude Haiku 4.5 leads in cost-effective JSON extraction benchmark
A recent benchmark evaluated six large language models on their ability to extract structured data, specifically JSON, from customer support emails. The analysis found that Anthropic's Claude Haiku 4.5 offered the best …
-
LLMs show promise for patient inquiry triage, but not autonomous deployment
Researchers have explored the use of few-shot large language models for categorizing online patient inquiries, aiming to improve clinical triage. They compared prompted LLMs against traditional methods like TF-IDF and B…