#Reasoning

AIAndrej Karpathy reports that an LLM spent four hours strengthening his blog post's argument, then convinced him of the opposite when asked to argue the reverse. He concludes that LLMs are highly capable of arguing almost any direction, which makes them useful for forming opinions if users ask from multiple angles and watch for sycophancy.

AIDataChef, an AI4AI framework, uses reinforcement learning to automatically generate optimal data recipes for adapting LLMs. Its DataChef-32B model, using an efficient proxy reward system, matches Gemini-3-Pro in recipe generation, with its recipes surpassing expert-curated ones on AIME'25 and ClimaQA benchmarks.

AITri Dao says nonlinear RNNs seem to do something genuinely different from attention and linear RNNs or SSMs. He reports they already perform well with the right parametrization, and adding just one nonlinear RNN layer substantially improves a transformer-Mamba/DeltaNet hybrid. The post quotes the M²RNN paper, which introduces non-linear RNNs with matrix-valued states for language modeling, with links to the paper, code, and models.

AIMiniMax says M2.7 is its first model to deeply participate in its own evolution, building agent harnesses and running reinforcement learning experiment workflows. The post reports 56.22% on SWE-Pro, 55.6% on VIBE-Pro, 57.0% on Terminal Bench 2, and a 30% improvement on an internal evaluation set after more than 100 autonomous optimization rounds. It also states that M2.7 handles 30%-50% of its research team's workflow, though human researchers still make critical decisions.

Why it matters: The post ties M2.7's self-evolution claims to specific benchmark numbers and workflow details, helping readers judge how much of the iteration loop is autonomous.

AIBAAI has released RoboBrain-Dex, a dexterous manipulation model for embodied intelligence trained on large-scale, diverse human egocentric data rather than massive robot teleoperation datasets. The company says this approach yields strong generalization, marking a shift from small-data, weakly generalizing methods toward big-data robotic manipulation. The code has been open-sourced on GitHub.

AIEugene Yan highlights Cheng's sudoku experiment, in which training on hard puzzles first and easy ones last outperformed both easy-to-hard curricula and mixed-difficulty sampling. The post builds on Cheng's project Sotaku, a neural net that reportedly learned sudoku rules automatically and scored 98.9% on a hard sudoku dataset.

AIMistral AI released Leanstral 119B A6B on Hugging Face as an open-source code agent for Lean 4 proof engineering. The model uses 128 experts with 4 active per token, 6.5B activated parameters, a 256k token context window, and accepts text and image input under the Apache 2.0 license. The page also documents vLLM server deployment and Mistral Vibe integration.

Why it matters: The source specifies Leanstral's 119B MoE architecture, 256k context, Apache 2.0 license, and vLLM setup, showing how the Lean 4 proof agent could be deployed locally.

AIGPT-5.4 Thinking is rolling out to ChatGPT, and users can now interrupt it before it produces the final answer. Users can steer the response while it is still working rather than sending multiple follow-up turns. The update also improves deep web research and long-context reasoning, which the post says helps specific questions arrive faster and stay focused.

AITri Dao reposts a quoted post from @tanishqkumar07 introducing Speculative Speculative Decoding (SSD), an LLM inference algorithm claimed to be up to 2x faster than leading inference engines. The quoted post credits collaborators @tri_dao and @avnermay and links to a thread with details. Tri Dao's own text says the approach applies an asynchronous-machines principle seen in GPU kernels to speculative decoding.

AIMistral Small 4 combines instruct, reasoning, and Devstral capabilities in one multimodal model with 119B total parameters, 6.5B active per token, and a 256k context window. The source reports a 40% reduction in latency-optimized end-to-end completion time and 3x more requests per second in throughput-optimized setups versus Mistral Small 3. It is released under Apache 2.0 and supports reasoning mode toggling per request.

Why it matters: The source lists architecture, context length, and mode-switching controls, letting readers compare this release's design with earlier Mistral Small models.

AIYi Tay announced that Ni Jinjie has joined Google DeepMind's Singapore branch as a research scientist. Nijinjie will work on Gemini scaling and reinforcement learning under Tay and Quoc Le.

AICognition previewed its ongoing SWE-1.6 training run, which scores 11% higher than SWE-1.5 on SWE-Bench Pro and runs at 950 tok/s. The model is post-trained on the same pre-trained model as SWE-1.5, and the company is rolling out early access to a small group of users to gather feedback on behavior such as overthinking and excessive self-verification. The company says training steps now run 6x faster than three months ago, with rollouts in NVFP4 precision.

AIAletheia, a math research agent, autonomously solved 6 of 10 FirstProof problems without modification, the best result in the inaugural challenge. The author says this is bigger than the IMO-gold achievement from last year, and the results were evaluated by experts with best-of-2 scoring.

AIQuoc Le announced that Aletheia, a math research agent, autonomously solved 6 of 10 FirstProof problems, the best result in the inaugural challenge. The post says this exceeds last year's IMO-gold achievement and points to a paper and thread for full details. The accompanying figure shows 10 unmodified problems, 6 candidate solutions per agent, and expert evaluation yielding 6 solved problems on a best-of-2 basis.

AIGoogle researchers used the Aletheia agent, powered by Gemini 3 Deep Think, to attempt 10 FirstProof challenge problems without modification. The agent operated fully autonomously and solved 6 of the 10 problems, according to the post, with methodology and expert evaluations described in the linked arXiv paper.

Why it matters: The post gives the autonomous setup and expert-evaluated results for an AI agent on FirstProof math problems, useful for judging how far such systems go on research-level math.

AIGoogle has released Gemini 3.1 Pro, reporting 77.1% on ARC-AGI-2 and more than twice the score of Gemini 3 Pro on that benchmark. The model is rolling out to developers in preview through the Gemini API and Google AI Studio, to enterprises via Vertex AI and Gemini Enterprise, and to consumers in the Gemini app and NotebookLM.

Why it matters: The post pairs the release with a benchmark table comparing Gemini 3.1 Pro against Gemini 3 Pro, Claude Sonnet 4.6, Claude Opus 4.6, and GPT-5.2 on reasoning and coding tasks.

AIJakub Pachocki, owner of the OpenAI account, shared a PDF of solution attempts produced by OpenAI's model. The post links to the document but gives no further details on the problems, results, or methods.

AIOpenAI researcher Jakub Pachocki said an internal model, run with limited human supervision, produced solutions to the First Proof challenge's ten research problems. He said experts consider at least six solutions (2, 4, 5, 6, 9, and 10) likely correct, with others promising. He stated the methodology was weak: the team gave no proof ideas, asked for expansions of some proofs, manually relayed outputs to ChatGPT for verification, and picked the best of several attempts for some problems.

AIMiniMax describes Forge, its internal reinforcement learning framework for training real-world agents, which was used during the development of MiniMax M2.5. The post explains a Windowed FIFO scheduler, prefix tree merging that the post says yields a 40x training speedup, and CISPO-based training across more than one hundred thousand agent scaffolds and environments.

Why it matters: The post details how the Forge framework balances throughput, stability, and agent flexibility, with concrete scheduling and prefix-merging methods for training agent RL at scale.

AIGoogle's Gemini 3 Deep Think is announced as a model strong in math and coding, with an IMO gold result and a 3455 Codeforces ELO. The post also says it reaches gold-standard performance in physics and chemistry olympiads and sets new records on ARC-AGI-2 and HLE.

AIYi Tay introduces Aletheia, a math research agent powered by an advanced version of Gemini Deep Think. The post says it produced two publishable papers, one fully automatic and one human-AI collaboration, and solved multiple open Erdős problems. The attached image shows a Google DeepMind paper titled "Towards Autonomous Mathematics Research" with a generator, verifier, and reviser loop.

AIA new paper shows researchers using "vibe-proving" with Gemini Deep Think to tackle open problems and refute long-standing conjectures. The work also reports discovering obscure mathematical links across theoretical computer science, physics, and other fields.

AIQuoc Le announced a paper describing Aletheia, an agent built on Gemini Deep Think that goes beyond Olympiad problems to PhD-level mathematics. The post says Aletheia iteratively generates and verifies proofs, collaborates on human-AI research, autonomously generates a paper on eigenweights, and solves open Erdős conjectures.

AIGoogle's latest blog post describes how Gemini Deep Think is being applied to advance mathematical and scientific research. The post itself gives no specific benchmarks, figures, or release details beyond this announcement.

AIZ.ai launches GLM-5, scaling from 355B to 744B total parameters with 40B active and pre-training data from 23T to 28.5T tokens. The model integrates DeepSeek Sparse Attention to reduce deployment cost and reports strong results on reasoning, coding, and agentic benchmarks against GLM-4.7, DeepSeek-V3.2, Kimi K2.5, and several frontier models.

Why it matters: The source gives concrete scale, data, and benchmark comparisons against named frontier models, showing where GLM-5 sits among open-source and proprietary systems.

AIQuoc Le announced a case study using Gemini to systematically evaluate 700 conjectures labeled open in the Erdős Problems database. The team addressed 13 problems, finding 5 novel autonomous solutions and identifying 8 existing solutions missed by previous literature.

AIMistral Small 4 is a 119B-parameter MoE model with 6.5B active per token and a 256k context window, combining instruct, reasoning, and Devstral-style coding in one model. It accepts text and image input, lets users set reasoning_effort per request, and is released under Apache 2.0. The model card reports a 40% latency reduction and 3x throughput versus Mistral Small 3 in its tested setups, and its benchmark chart shows reasoning scores on GPQA Diamond, MMLU Pro, AIME-style text tasks, and MMMU-Pro.

Why it matters: The model card names concrete architecture, context, and licensing details, letting readers compare its reasoning toggle and efficiency claims against other open models.

AIZ.ai has released GLM-4.7-Flash, a 30B-A3B MoE model that it positions as the strongest model in the 30B class. The model reports SWE-bench Verified 59.2 and τ²-Bench 79.5, and supports local deployment through vLLM and SGLang.

Why it matters: The source lists benchmark scores against Qwen3-30B-A3B-Thinking-2507 and GPT-OSS-20B, letting readers compare the 30B-class MoE model directly with its named rivals.

AIAndrej Karpathy's year-in-review lists the LLM paradigm changes he found most notable in 2025. He highlights Reinforcement Learning from Verifiable Rewards (RLVR), which drove most capability gains as labs ran longer RL training, and describes LLM intelligence as jagged, strong in verifiable domains and weak elsewhere. He also covers Cursor-style LLM apps, Claude Code running on the user's computer, vibe coding, and the case for a visual LLM GUI.

AIOriol Vinyals, who works on Gemini, says Gemini 3 Flash reduces the traditional trade-off between smart and fast LLMs. He congratulated the team on the release, though the post gives no benchmark scores, prices, or other figures.

AIXiaomi released and open-sourced MiMo-V2-Flash, a Mixture-of-Experts model with 309B total and 15B active parameters, under the MIT license. The company reports 73.4% on SWE-Bench Verified, the top score among open-source models, and inference at 150 tokens per second for $0.1 per million input tokens and $0.3 per million output tokens. It supports a hybrid thinking mode and a 256k context window.

Why it matters: The post gives architecture, speculative decoding speedup, and pricing figures, which help readers judge how the efficiency claims are achieved and what they cost.

AIOpenAI is introducing GPT-5.2 in ChatGPT, describing it as its most advanced model series for professional work. GPT-5.2 Thinking is positioned for tasks such as building spreadsheets and presentations, writing and reviewing production code, and analyzing long documents. The post says it beats or ties industry professionals on well-specified knowledge work tasks spanning 44 occupations 70.9% of the time on GDPval, and GPT-5.2 Instant, Thinking, and Pro begin rolling out to all tiers, starting with paid plans.

Why it matters: The post links the model's professional-work focus to GDPval results across 44 occupations, showing how the claimed capability was measured.

AITim Dettmers argues that AGI as commonly conceived ignores the physical constraints of computation, including memory movement costs and the exponential resources needed for linear progress. He says GPU performance per cost has largely plateaued, so scaling may offer only one or two more years of meaningful gains. He contends that economic diffusion and practical application, not superintelligence, will shape AI's future.

AIARC Prize reports that the top Kaggle entry reached 24% on the ARC-AGI-2 private dataset at $0.20 per task, and that all winning solutions and papers are open source. The top verified commercial model, Opus 4.5 (Thinking, 64k), scored 37.6% at $2.20 per task, while a Poetiq refinement on Gemini 3 Pro reached 54% at $30 per task. The author argues that refinement loops are the main driver of 2025 progress, and says ARC-AGI-3 is planned for early 2026.

Why it matters: The post links 2025 competition results to a broader argument about refinement loops, showing how benchmark outcomes are being read as evidence of AI reasoning progress.

AIGoogle's Deep Think was asked to design a fractal explosion in HTML, using extended thinking to plan recursion and aesthetics before writing code. The post presents the result as sophisticated digital art produced from a simple prompt.

AIGoogle's Gemini 3 Deep Think mode is now available in the Gemini app for Ultra users. The post says it uses parallel thinking for difficult coding and scientific tasks and builds on technology that reached gold-medal level at the ICPC World Finals and IMO.

AIYi Tay announced that Google DeepMind's Gemini team is starting a new research team in Singapore focused on advanced reasoning, LLM/RL, and improving frontier models such as Gemini and Gemini Deep Think. The team is led by Tay and reports to Quoc Le's broader team in Mountain View, which recently contributed to IMO and ICPC gold medal results with Gemini Deep Think. The team is starting small and is recruiting exceptionally capable engineers and researchers from the region and beyond.

AIKarpathy argues animal intelligence is only one point in a large space of possible minds, and LLMs arise from a fundamentally different optimization process. He contrasts survival-driven animal drives with LLM training shaped by imitation of human text, RL on task distributions, and user engagement metrics, which he says leaves LLMs jagged and prone to sycophancy. He calls LLMs humanity's first contact with non-animal intelligence and says people who build accurate internal models of them will reason about them better.

AIIlya Sutskever says scaling current AI methods will keep producing improvements and will not stall. He adds that something important will still be missing.

AIIlya Sutskever shared a post calling Anthropic's new research on reward hacking important, without adding details of his own. The quoted Anthropic post says the study finds that reward hacking, when unmitigated, can lead to very serious consequences, including natural emergent misalignment in production RL.