Open source

AIInferact, working with DeepSeek, NVIDIA, and SemiAnalysis alongside the vLLM community, says joint work across models, custom kernels, and engine serving cuts time to first token (TTFT) by nearly 70% at ~100K throughput. vLLM is the open-source inference engine, and Inferact optimizes it for enterprise production deployments.

AIA solo founder has built vaultinspo.com, a free AI image library where every image comes with the exact prompt that created it. Users can copy and remix the prompts in tools such as Midjourney, Flux, or SDXL, with no login or paywall required.

AIOpenAI announced on X that it is releasing a series of new mathematical results produced by its internal frontier model. The material is reported as 722 manuscripts posted to GitHub, and the excerpt provides no further detail on the specific results.

AIDeveloper Brandon Thomas used Claude Opus 5.5 and Rust to build Artcraft, a free open-source suite with Photocraft, Vectorcraft, and other apps that mimic Adobe products. The author tested Photocraft and found its basic editing commands worked where expected, but Free Transform behaved unpredictably. Thomas describes the software as early alpha and invites developers to contribute.

AIFrench AI company Mistral has released Mistral Large 4, its new flagship model with 1 trillion total parameters. The company plans to make all of the model's weights open at the end of the month.

AIGoogle's AI Edge Team open-sourced ML Drift under Apache 2.0, a GPU compute engine for on-device AI inference across OpenGL ES, OpenCL, Metal, and WebGPU. It serves as the core GPU acceleration engine within LiteRT and succeeds the legacy TFLite GPU delegate, which will no longer receive new features. The post cites benchmarks showing up to 40% lower frame latency in YouTube Shorts and up to 30% faster on-device performance in Adobe Lightroom and Photoshop.

Why it matters: The post explains how ML Drift unifies GPU shaders across platforms and replaces the TFLite GPU delegate, which matters for developers deploying on-device models.

AIEpoch gave six models 11 real work tasks from its own operations, including graphic design, data insights, and research design, and graded outputs against employee standards. Fable 5.1 and GPT-6 Astra led on average task performance, reliably handling well-defined work such as coding and computational analysis. The report finds that all models still fail on open-ended judgment, including matching Epoch's standards, designing informative experiments, and generating diverse ideas, so the authors conclude AI cannot yet replace workers at Epoch.

Why it matters: The report separates well-defined task reliability from open-ended judgment failures, which benchmark scores on easily verifiable tasks would miss.

AIA Hugging Face blog author used the ML-Intern agent in HuggingChat to build six small models by writing detailed prompts that specify datasets, base models, baselines, smoke tests, and spending limits. The projects include a citrus disease vision-language model, a Huggy character LoRA, a camera-angle LoRA, a doodle-to-object LoRA, a 0.8B prompt rewriter, and a 4-step distilled Agate model, with total compute cost of about USD 103. Each project's prompts and public models are linked from the post.

Why it matters: The author shows how prompt structure, baselines, smoke tests, and budget caps shape an agent-driven training workflow, with per-project costs given.

AITeknium, a Nous Research affiliate, posted only the word "hello" on X. The quoted post from Nous Research announces a Series B raise to advance Hermes Agent and build a mobile app, with investors including NVIDIA and Samsung Next.

AILiquid AI's local d1 3B model, running on a single RTX 3090, made decisions in 8 ms and ate 86 apples in a Snake game, versus 3 apples for OpenAI's API, which took 345 ms per call. The post, shared via Atomic Chat, also credits the d1 model with 82 moves per second.

AIDeveloper Brandon Thomas's Artcraft has launched seven open source apps in Rust that aim to replicate the interfaces and tools of Adobe Photoshop, Illustrator, Premiere, Lightroom, After Effects, InDesign, and Acrobat Pro. Thomas said he used Anthropic's Claude Opus 5.5 to build the clean-room replacements, with WebAssembly versions available for browser use. The apps remain in a "super early alpha" state, and commenters have pointed out many current shortcomings.

AIPerplexity's open-weights multimodal decision model, pplx-decider-v1.1-27b, is now available on OpenRouter. It accepts text, JSON, or images and returns typed answers with probabilities, priced at $0.02 per million input tokens with free output.

AITogether AI, IBM, and Nvidia are scaling open-source inference on a dedicated B300 cluster hosted on IBM Cloud. IBM Cloud GM Alan Peacock argues that spending on tokens is worthwhile when the business benefits from them.

AISnorkel AI is expanding its Open Benchmarks Grants tenfold to a $30M commitment to fund more diverse, robust, and continuously updated open AI benchmarks. The program adds an Open Benchmarks Red Team to test and strengthen those benchmarks, plus a Snorkel Research Fellowship for independent researchers developing new evaluation methods. The source says OBG-funded benchmarks have appeared on model cards from every major frontier lab.

AIPerplexity's open-weights multimodal decision model, Decider v1.1, is now available on OpenRouter. It accepts text, JSON, or images and returns typed answers with probabilities, priced at $0.02 per million input tokens with free output. Perplexity says it scores highest on Hugging Face's new Decision Index 0.3 benchmark and costs half as much as v1.

AITogether with DeepSeek's model and kernels and NVIDIA's collaboration, vLLM reports that time to first token (TTFT) drops nearly 70% at roughly 100K throughput. The post credits Inferact and the vLLM community for the work and thanks SemiAnalysis for AgentX.

AIvLLM has integrated DeepSeek's MegaAttention, which uses NVFP4 KV cache and is 45% smaller, alongside Mega-mHC, Mega-Gate, and DeepSelect kernels. Its own fusions include a CuTe-DSL fused WO-A that lowers ITL by up to about 6–7%, and sparse MQA logits that run 14–23× faster per layer at 512K context.

AIvLLM's bounded sliding-window attention (SWA) replay reruns only each request's last 128 tokens after a prefix hit, instead of replaying 40 × 128 tokens. With CUDA graphs, prefill compute drops 30–40% for layers 21–39.

AIThree weeks after launch, vLLM's serving of DeepSeek-V4.1-Flash runs 1.9× faster at low concurrency and delivers 5.3× the throughput at 150 TPS per user on SemiAnalysis AgentX. The post offers interactive figures explaining how these gains were achieved.

AIAnthropic's ClaudeDevs account recommends compatible browser drivers from Browser Use, Browserbase, E2B, and Daytona for computer-use tooling. Developers can also write their own driver based on the example drivers in the quickstart. Quickstart and documentation links are provided for the computer-toolset setup.

AISam Altman reposted an OpenAI announcement that GPT-6 and Intelligent UI are rolling out in ChatGPT for everyone. According to the quoted post, Intelligent UI produces fast, interactive answers with visual explanations and on-the-spot tools for tasks.

AIAWS has open sourced Strands Box, a local sandbox for AI agents. The author points readers to a blog post by Marc Brooker that covers design tradeoffs, how OS-level containment and Dogwood enable richer policies, and AWS's investment in agent safety.

AILocal sandboxing for GitHub Copilot is now generally available. It lets Copilot run commands in an isolated environment with controlled access to files, networks, system capabilities, and credentials. Enterprise teams can also centrally manage policies, and the feature is available in GitHub Copilot CLI, the GitHub Copilot app, and @code.

AIOpenAI Developers highlighted a hockey game built in a single shot with Codex, shared alongside the post. The game, made during DevDay and Edmonton Oilers opening night, has McDavid feeding Draisaitl one-timers.

AIIdeaLens is a detector that identifies whether the ideas in a text came from a human or an AI, rather than judging the prose alone. On mixed-provenance benchmarks, it reached 81.3% average idea-detection accuracy, versus 25.4% for ProseLens and 25.9% for Pangram 4. The model, code, and data are open-sourced, and it was trained on Tinker.

AICline says users can try its tool for free with models such as DeepSeek-V4.1-Flash. It also offers ClinePass, a subscription for open-weights models at roughly 5x discounted access.

AIGoogle released EmbeddingGemma 2 on Hugging Face, and Paige Bailey called it a step toward open models for open science. The post links the model and cites earlier EmbeddingGemma-based projects, including medical, geographic, oncology, and PubMed embedding models.

AILiquid AI released two open-weight multimodal decision models, d1-3B and d1-omni-600M, which return probability answers in one forward pass with zero output tokens. d1-3B scores 48.57 on Decision Index v0.2.1 and answers one question in 8 ms on an RTX 4090, while the models are licensed free for commercial use below $10 million in annual revenue.

AICloudflare's open-source Clef (27B) and Clef Flash (9B) decision models are available on OpenRouter. They accept text, JSON, or images and return typed answers with probabilities rather than generated text. Pricing is $0.24 per M input tokens for Clef and $0.09 per M for Clef Flash, with output free.

AIClaude Code v2.1.293 adds Claude Haiku 5.5 (claude-haiku-5-5), now the default Haiku model on the Anthropic API, with 1M context and pricing of $0.10/$0.50 per Mtok ($0.50/$2.50 for prompts over 100K). The release also adds agentType to the subagentStatusLine payload and isDeferred to $.tool.register, and fixes numerous issues including a memory leak in HTTP MCP connections.

AIPreference Model is open-sourcing Karotte, the framework it uses to build reinforcement learning environments that resist reward hacking, including defenses like killing stray processes before grading and rejecting grader-crashing files. The framework has been hardened through more than a million evaluation runs and controlled red-teaming. The company focuses on machine learning engineering tasks for leading labs, and a16z says it is partnering with Preference Model and its founders, Jennifer Zhou and Ning Cao.

AIllama.cpp can distribute inference across heterogeneous devices through the ggml RPC backend, according to Georgi Gerganov. He says it is currently an advanced setting, but he expects it to become more accessible to regular users over time. A related post reports MiMo 2.6 Flash running across an RTX 6000 GPU and an M5 laptop over 10 GbE at about 40 tokens/sec.

AILiquid AI and NVIDIA Robotics announced a collaboration around Liquid AI's models on NVIDIA hardware. The main post is brief and gives no further technical detail or figures.

AIUnsloth says users can train their own Decision Model on as little as 2.5GB of VRAM using small models such as Laya. Training is done through a UI, with a video tutorial and guide linked from the post and the project's GitHub repository.

AIAn open-source GitHub repository called AI Engineer Headquarters offers a structured curriculum for aspiring AI engineers. According to the post, it covers fundamentals, large language models, RAG, fine-tuning, tool calling, and agentic workflows.