Tech News
← Home  ·  All topics

Gpt

250 GoKawiil briefs on this topic · RSS for gpt

Open-weight models' share of LLM Atlas price-performance frontier falls to 3%

Village's LLM Atlas, which tracks the strongest AI model at every price point, now shows open-weight models winning only 3% of budget tiers, down from 41% in June. Most of that decline followed OpenAI's GPT-6 Luna and GPT-6.1 Terra releases along with Anthropic's Claude Haiku 5.5, which reclaimed budgets previously held by DeepSeek V4 Flash and Zhipu's GLM-5.2.

Blogger claims local Qwen 3.8 27B model can read and edit neural network weights

A developer reports replicating findings from a paper titled 'Sparks of Weight-Space Programming,' which described GPT-6 Astra analyzing and editing raw neural network weights directly from text. The blogger says a locally run Qwen 3.8 27B model performed similar weight interpretation tasks on a laptop, following an example from the paper.

MIT team uses GPT-6 Astra to directly edit Qwen model weights, lifting math benchmark scores

MIT researchers reportedly applied OpenAI's GPT-6 Astra system to modify the open weights of Alibaba's Qwen model directly, rather than through standard fine-tuning. The intervention is said to have improved the model's performance on math-related benchmarks.

Report disputes claims that 'Astra' AI reads raw weights to improve Qwen model

A critical analysis challenges a widely shared post claiming that an AI system called Astra (described as GPT-6 Astra) can improve neural networks purely by inspecting raw weights, including boosting Qwen3.5-0.8B's math performance from 7.9% to 48.1% on MATH-500. The analysis finds that only small-scale experiments, like naming what a PPO-trained network does, actually work by reading raw weights alone; the Qwen math-improvement experiment and a game-playing network experiment relied on running the network and using Python tools rather than weight inspection alone.

OpenAI expands ChatGPT Dots with GPT-6 Astra and GPT-6.1 Terra models

OpenAI is adding new AI models, including GPT-6 Astra and GPT-6.1 Terra, to power ChatGPT Dots, its persistent personal AI assistants. According to OpenAI's Thibault Sottiaux, Dots will now automatically select the best model for a given task and can switch between models mid-task without user input. Terra is said to offer faster responses and a friendlier tone, and will help OpenAI roll out Dots to a wider audience.

Study: GPT-4o grant screening favored conventional proposals over novel ones

Diogo Machado of consultancy Technopolis tested OpenAI's GPT-4o on 139 real grant proposals submitted to Austria's Science Fund in 2021, asking it to pick the better of each pair. The model matched human reviewers' funding decisions 60% of the time, but was notably less likely than humans to back proposals rated high for novelty, according to the study published in Scientometrics.

Astra AI system edits Qwen3.5-0.8B's raw weights, lifting MATH-500 score sixfold

A research team reports that its frontier model, called Astra, can directly read and edit neural network weights without any training loop or gradient updates. By identifying a reasoning-loop flaw in Qwen3.5-0.8B's weights, Astra edited roughly 0.016% of them and raised the model's MATH-500 score from 7.9% to 48.1%, with IFEval scores tripling. Astra also reportedly wrote small game-playing networks from scratch that beat networks trained with PPO over 100 million steps, and could identify a network's training task just by inspecting its weights.

Study: 50 GPT-powered agents built bureaucracy, not civilization, in Minecraft test

Researchers placed 50 AI agents, each running an OpenAI GPT-6.1 Sol model, into Minecraft for a three-hour trial aimed at building a 'thriving civilization,' using an open-sourced agent harness built for the experiment. Instead of cooperating efficiently, the agents established leadership purely by directory order, invented redundant administrative roles like 'rod holder' and 'fishing shift,' and spent long stretches on trivial coordination tasks while failing to secure a stable food supply.

OpenAI details safety tuning for GPT-Rosalind-6 biology research model

OpenAI published a system card addendum describing GPT-Rosalind-6, a model built to support advanced life sciences research, including some dual-use topics that its general-access GPT-6 Astra model would refuse. The company says the model is separately trained to refuse or safely handle requests that could meaningfully enable severe biological weapons development, with this training kept distinct from its classifiers and offline mitigations. OpenAI tracks refusal quality using held-out synthetic data, red-teaming, and production prompts, noting its benchmarks and grading methods are still being updated.

Census Bureau survey wording change inflates reported AI adoption among large firms

Analysis of the U.S. Census Bureau's Business Trends and Outlook Survey shows that large companies' reported AI use jumped from about 12% in September 2025 to roughly 40% by October 2026. The increase largely reflects a change in the survey question, from asking whether AI was used in 'producing goods or services' to asking about use in 'any business function,' rather than a sudden surge in actual adoption.

IEEE's Review Process Catches AI-Generated Fake Papers, Writer Finds

Freelance writer Benjamin Skuse tested IEEE's peer-review defenses by submitting AI-generated scientific papers, in one case simply running an automated 'run experiment' and 'write paper' process with no edits before submission. According to IEEE features editor Eliza Strickland, the combination of human reviewers and research-integrity software currently managed to catch these machine-made submissions.

Y Combinator and Olam Labs test AI agent swarms in Minecraft civilization experiment

Researchers from Y Combinator and Olam Labs ran experiments putting swarms of AI agents into Minecraft, including a 'Civ50' test where 50 GPT-6.1 Sol agents were given three hours to build a thriving civilization. Instead of flourishing, the colony stalled after zombie attacks thinned their numbers, with agents abandoning ambitious plans and spending long stretches on trivial bureaucratic tasks, such as four agents taking 11 minutes to appoint a 'sole fertilizer operator' for one piece of bone meal. The team also tested adversarial multi-agent scenarios, including the PvP mode Bed Wars.