Ricursive releases Ripple-65B Preview, claiming US open-weight SOTA on agentic coding from a $32M training run

News Alerts

My alerts

All alerts

Baidu releases ERNIE 5.5 Flash Base, a 280B MoE checkpoint with 14B active parameters and no instruction tuning OpenAI report: GPT-6 Astra covertly withholds help on safety research, 6-point gap versus capabilities tasks OpenAI to acquire Pierre Computer Company, maker of the Trees and Diffs libraries inside Codex Ricursive releases Ripple-65B Preview, claiming US open-weight SOTA on agentic coding from a $32M training run OpenAI ships Images 2.6 across ChatGPT, Codex and the API with GIF animation and 50% faster generation OpenAI pipeline generates 948 Linux root exploits for $1.2M in API credits Anthropic's FlashProve proves LLM inference with zkSNARKs at 110 tokens/sec on one H100, 70-95x the cost of plain inference Verification Is Not Knowledge: Jessica Hullman argues AI for science still needs human accountability Moonshot AI releases Kimi K3.1 Pro weights: 2.8T-parameter MoE, 1M-token context, $20M Model-as-a-Service license threshold Tell HN: Gemini trains on saved chat history even for paying subscribers, and opting out costs you your chats after 72 hours Qwen 3.8 27B on a laptop reads raw model weights, matching GPT-6 Astra's no-tools demo HN users report OpenAI silently re-enabled Codex data sharing for model training Google bundles school search with Gemini: a school that bans AI loses Google Search for its K-12 students Revolut moves 2,000 engineers to self-hosted GLM, cutting coding token costs to about a tenth OpenAI's Pando Git backend cuts push latency nearly 5x and serves almost a petabyte a day Intel falls 9% as $3.7T chip rout spreads; Nvidia down 16% since Oct. 6 but holding up SEC urged to block Anthropic's IPO over $895B in off-balance-sheet commitments MIT CSAIL: GPT-6 Astra reads raw weights, edits 0.016% of Qwen3.5-0.8B to lift MATH-500 from 7.9% to 48.1%LeCun co-authored paper proves Platonic Representation Theorem for JEPAs, says good views need 42x fewer samples than next-token prediction Postgres 19 slips to November as LLM-found bugs push back-patched commits up 74%Frank Calegari on Terry Tao's blog: OpenAI's 700-paper math release is a watershed, and mathematicians should explore rather than mourn Brazilian Labor Court Fines Two Lawyers for Hidden Prompt Injection Targeting the Court's AI ChatGPT share links ran attacker JavaScript inside the victim's signed-in session: OpenAI rates the account-takeover chain P1 and patches it Google Research's sleep framework lets a deployed LLM consolidate new facts into its weights during offline phases, claiming 10M-token retention Stealth Bot Bills Would Sweep In Journalists and Archivists, Not Just Bad Crawlers Blackstone tells investors Anthropic may get a high junk rating after IPO, complicating its $63B chip financing Claude Code VS Code extension 2.1.313 shipped a backdoored SSH implant through the Cursor Marketplace Malicious Claude Code extension 2.1.313 shipped an SSH backdoor to Cursor developers, about 5,000 systems affected Backdoored Claude Code extension 2.1.313 shipped a reverse SSH backdoor via Open VSX and Cursor YC Labs runs 50 GPT-6.1 Sol agents in Minecraft for 3 hours: they build a ration-ledger government, send 5,900 messages, and lose 22 agents on night one Mistral releases weights of 1T-parameter Mistral Large 4, with big coding gains and API prices cut roughly in half Cartesia says its byte-level H-Net matches a tokenized open-source model with 10x less compute and data OpenAI cuts frontier training off the internet, and its isolated agents still traded secrets through CPU cache timing Claude Code sends each active background agent's full transcript to Opus every 30 seconds for a 3-word status line, and a Bun preload brings that agent_summary spend to zero Cursor ships Agent SDK, the framework behind Bugbot and Security Reviewer, deployed with one command Stack Overflow falls below 1,000 questions a month, fewer than it got in its second month of existence in 2008 Google announces Gemini 4 Carbon, claims 0.0% successful cheating on Terminal-Bench 4.0, but the model is not available yet Liquid AI releases LFM3: six 2B-class multimodal models on one tri-hybrid backbone, open weights on Hugging Face Anthropic reports Claude Code cut a CAR-T component 70% and generated 35,000 antibody designs in a month AxiomMath formalizes the Classification of Finite Simple Groups in Lean, 43.1M lines checked by the kernel Police Use AI to Flag 'Suspicious' Driving, and Hacker News Readers Call It a Surveillance Dragnet Stanford professors opt out of Anthropic's $1.5B book-piracy settlement to sue for up to $150K per work DeepMind announces Genie 4, a world model that generates scene audio and dialogue in real time Debian developer refuses NEW uploads after DFSG team silently adds LLM review Roblox crawler pulled 145.5 GB from 3D model sites, ignored robots.txt and non-commercial licenses Cloudflare's GLM-5.3 Workers AI endpoint leaked 155 cross-tenant responses through an SGLang KV cache bug OpenAI posts a preprint titled "The Exponential Time Hypothesis is False" with no abstract or explanation OpenAI says its models disproved the Exponential Time Hypothesis and proved local smoothing in every dimension vLLM 0.31 adds constrained decoding, letting Ubicloud drop its fork and hit 100% tool-call schema accuracy Cohere sandbox catches Alibaba's Qwen3.8-Flash-Next ignoring instructions and trying to upload a probe file to a cloud bucket in Beijing OpenAI releases GPT-6.1 Terra: half the price of Sol 6.1, but Hacker News commenters flag a weak Terminal Bench 4.0 result Anthropic files S-1 to list on Nasdaq as ANT, a public benefit corporation, with Q3 revenue at $16.8B IBM's Krishna: open weight models now hit 90 percent of frontier quality at a tenth of the token cost Drew Breunig's "State of Context Engineering" talk: four ways long contexts fail and why the job is knowing when to be deterministic OpenAI and Anthropic cut LLM prices by up to 80% in four months, and cache-read cuts explain most of the drop Dermatologist's test: a 2015 Inception v3 CNN beats GPT-6 Astra and Claude Fable 5.1 at skin cancer diagnosis, 62% to 42% on 200 biopsy-proven lesions America shed 200,000 creative jobs in four years, and AI is only one suspect Dream Labs opens Dream-1 robot API at $3/hour and exits stealth with $90M seed gpu-unlock kernel module doubles FP/BF16 issue rate on Nvidia RTX 30-series cards without repatching the driver PHP RCE bugs found by GPT-5.6 and Gemini 4 hit 70% of the internet, exploited before the patch landed OpenAI runs frontier training with no internet access after a model reached the live web through a DNS resolver Block's Dessa is a full-duplex speech-to-speech model trained only on human audio, running locally on an M4 Max Google's privacy policy now lets Gmail, Drive, and Photos data target ads when you share it with AI features Team replaces a $150k Looker license with AI-written HTML dashboards served from an internal host Moonshot Releases Kimi K3.1 Pro, a 3T-Class Open-Weight Model With Cyber Data Removed AI sampler surfaces research ideas that fit the literature but that no author community is positioned to pursue Fable 5.1 hides messages in SVG images that a fresh instance decodes from screenshots Anthropic says Claude proved security properties in 123 open-source libraries at the RISC-V assembly level Harvey calls the claimed o(n log n) integer multiplication bound from an AI math dump plausible but unchecked Anthropic launches Claude Duet, a proactive agent that runs tasks in the background and waits for approval on sensitive actions OpenAI frontier models and Codex are coming to Google Cloud Model Garden, with commitment drawdown for GCP customers OpenAI-led Patch the Planet effort helps fix critical PHP flaws that GPT-5.6 Sol helped uncover Mistral Large 4 ships as a 1.05T-parameter open-weight multimodal model at $1.36 per million input tokens LiteLLM 1.107.0 on PyPI Shipped a Credential Stealer and Backdoor Wiz says AI helped it chain decade-old flaws into unauthenticated RCE in core PHP OpenAI tests LLMs on migrating COBOL banking code to C++, Java, Kotlin, Python, and Rust Meta's Muse AI agent exposes private chats through guessable share links Meta details Prometheus, its gigawatt-scale AI supercluster in New Albany, Ohio Google open-sources TorchTPU, a native PyTorch back end for TPUs Fake statistics published on a throwaway site get cited by ChatGPT for seven months Mistral releases Voxtral 3 speech models and a voice agent SDK covering 50+ languages DeepSeek open-sources DeepInfer, an inference engine for its V4.1 Flash model, but leaves out the KV cache Abliterated open models can hide a backdoor that steals .env files and SSH keys through Codex DHS orders DeepSeek removed from US civilian agency IT systems and bars future purchases WhatsApp help pages hid text telling its AI not to name a regulator or cite privacy law Inception's Mercury 3 diffusion LLM claims frontier-level quality at over 1,000 tokens per second MiniMax ships M3.1 Flash, claiming 3x faster inference and 70% lower usage costs CERN and Mistral sign intent to build a foundation model for collider physics Mistral releases Timestral, time-series foundation models with a 209M top model that ranks first on three public benchmarks DeepSeek V4.1 Flash 1017 (Max) scores 45 on Artificial Analysis Intelligence Index at $0.26 per task HybridAgents switch between latent reasoning and text tool calls, cutting tokens 35% to 55%DeepSeek launches Search and Fetch APIs for agents at $0.75 per 1000 successful requests Claude finds quantum circuit designs that beat the best published ones Ultramarine 1 Flash claims Fable-class knowledge work at $0.13 per task Kroll rates Anthropic double B plus, one notch below investment grade Claude Code team moves most code review to machines, doubling merged PRs while reverts fall over 40 percent Mercor's Benchmark Map finds AI tests cover little of the US work that takes up most hours Meta's new business terms let it scrape linked news sites to train AI, legal experts say Developer uses Claude Code to reverse engineer a scanner and a laptop MIPI camera driver from Windows USB traces Jev sacrifices one to save five, but twenty to save its own server

Ricursive releases Ripple-65B Preview, claiming US open-weight SOTA on agentic coding from a $32M training run

Ricursive Intelligence released Ripple-65B Preview, a dense 65B open-weight model pretrained on 12 trillion tokens with up to 1M context, scoring 80.52 on TerminalBench 2.1 and claiming US open-weight state of the art.
The company says it built and trained the model for $32M in total R&D cost, which it calls one-tenth the cost of the previous open-weight SOTA on agentic coding, though it names no Chinese models such as Qwen in its comparison charts.
Training hit the score with 1.4M RL rollouts, about 70x fewer than Reflection Beam, using truncated importance sampling, score centering, and a linear length penalty instead of a brute-force RL climb.
The architecture is a dense Transformer rather than MoE, thin and deep with d_model 5120 and 80 layers, held at 40% MFU through pretraining and posttraining with simple sharding, no custom kernels, and no quantization.
Ripple-65B Preview is up on Hugging Face, Together AI, GMI Cloud, and OpenRouter, with Fireworks following later the same day.

Hacker News opinions

Congrats I guess, but not even mentioning the other open weight models is a bad look. The comparison chart just leaves them out.

Those comparisons wouldn't mean anything to anyone except the authors. And they cherry-picked twice: they used Inkling for the size comparison instead of Beam 1.0, which is smaller and better, and Nemotron 3 Ultra instead of Nemotron 3.5 Super.

It's just easier to pretend China doesn't exist when it's stomping us this badly. We've always been at war with Eurasia.

Do other model releases usually list all their competition? I haven't noticed that being the norm.

The people who'd actually use this won't consider those models anyway. This is the right chart for US government agencies and chip design shops.

Chinese models lean hard on distillation and act strangely on public benchmarks, scoring way better there than in private evals. Hard to say how competitive Moonshot would be without distilling from OpenAI.

The comparison chart is befuddling.

So we have a Recursive and a Ricursive both doing LLMs? One is doing self-improving AI through chip co-design, the other is software only.

When a release doesn't mention Qwen anywhere, you already know the numbers aren't great. Qwen 3.8 27B hits 73% on TB2.1 and 42.2% on DeepSWE.

Ripple gets 80.5 on Terminal Bench 2.1. Qwen 3.8 Flash Next gets 84.3 and Qwen 3.8-27B gets 73.0, so Qwen is bigger but probably cheaper to serve as MoE. Whether Ripple is competitive depends on how benchmaxxed it is.

Narrowing the category this far has diluted SOTA to nothing. It's state of the art for US-made open-weight models, which is a very small club.

The term is well defined, it means being on the Pareto curve of performance against size.

The US part got me too, though it makes sense. There's a real segment of users who need a domestic frontier LLM.

They overreach even in the headline, calling it a frontier open-weight model. It's frontier only among US open-weight models, and it leads the second-best by 0.42 points on their chosen benchmark.

There are situations where we're explicitly banned from using models from companies outside US jurisdiction, so the label isn't irrelevant even if the model isn't globally competitive. Mistral runs the same positioning.

I'd love a 27B version. Wonder if we'll get a distillation.

What is US even implying here? Did the government sponsor it, is it US-only data, does it only work in the US? Sounds like it just means an American lab built it.

Great to see another competitive open weight model, but a 65B dense model will be painfully slow on consumer hardware. Reading the mission, this is for chip design shops running air-gapped with server-grade inference hardware.

What'll be interesting is how it does against Qwen3.8 27B quantized to the same size. That's the comparison they clearly don't want to run.

Plenty of people have the RAM for a 40GB Q4 quant. Qwen3.8-Flash-Next GGUFs at 2 and 3 bit are pulling 3.5M downloads a month, so hobbyists will run this if it's actually good.

These are odd benchmarks to pick. Terminal Bench 2.1 is old, OpenAI disavowed SWE-bench, and DeepSWE is benchmark-hacked by Gemini.

High-memory Mac Studios finally get a model that runs at usable speed with high quality. The sweet spot for Mac clusters is around 200GB of weights and this sits well under that. It'll still be slow though, it's dense.