Qwen3-Coder 480B
PulseAugur coverage of Qwen3-Coder 480B — every cluster mentioning Qwen3-Coder 480B across labs, papers, and developer communities, ranked by signal.
- 2026-07-19 product_launch Qwen3 Coder 480B and A-35B Vengeance models have seen their prices drop by 44% to $1 per 1M output tokens. source
3 day(s) with sentiment data
-
New POLCA system boosts LLM serving efficiency by 20% · 2 sources tracked
Researchers have developed a new power control system for disaggregated LLM serving that optimizes GPU energy efficiency. This system, called POLCA, decouples power management for prefill and decode phases, unlike NVIDI…
-
LU Labs offers uncensored text AI models, clarifies image/video restrictions
LU Labs offers a free, open-source desktop application for Windows and Linux that allows users to run large language models locally on their own machines. The company emphasizes that their hosted text models do not have…
-
LU Labs offers hosted AI models as Mac LM Studio alternative
LU Labs has launched a hosted cloud service as an alternative to LM Studio for Mac users, particularly those with lower-spec machines or limited memory. The service offers access to a wide array of chat, image, and vide…
-
Hermes 3 405B model now available via LU Labs hosted service
LU Labs has launched a hosted service for the Hermes 3 405B model, a large language model requiring significant computational resources. The service is accessible via a web-based studio or through an OpenAI-compatible A…
-
OpenAI claims math breakthrough amid plagiarism accusations; new AI models released
OpenAI has reportedly achieved a significant breakthrough in mathematics, solving a long-standing open problem using a large number of AI agents. However, this achievement is overshadowed by accusations of intellectual …
-
OpenRouter's free AI models offer coding help with strict limits
The free tier on OpenRouter offers access to several AI models for programming tasks, but comes with significant limitations. Users are restricted to 20 requests per minute and a daily cap of 50 requests for new account…
-
WebGrader trains LLMs for web development with self-evolving grader
Researchers have introduced WebGrader, a novel system designed to train large language models for web development tasks. This self-evolving programmatic grader autonomously generates interaction flows from website reque…
-
Open-weight AI models see price drops and leaderboard additions
Several open-weight AI models have recently been added to tracking leaderboards or seen significant price reductions. Inkling Small from Thinkingmachines is noted for its large context window and competitive pricing, wh…
-
Chinese AI models undercut Western pricing by up to 50x, offering competitive performance · 4 sources tracked
A comparison of AI API pricing in 2026 reveals that Chinese providers like Zhipu AI, Baidu, DeepSeek, and Alibaba Group offer significantly lower costs than Western counterparts such as OpenAI, Anthropic, and Google. Mo…
-
Developer uses 9 AI agents for cost-effective coding workflow
A principal architect has developed a cost-effective coding workflow using a multi-agent system that leverages different AI models for specific tasks. This approach, called Spec-Driven Development (SDD), assigns distinc…
-
LLMs improved for power system code generation with new intervention
Researchers have developed a new method to improve the reliability of large language models (LLMs) for power system code generation, particularly for on-premise deployments. The approach addresses API knowledge boundary…