LLM News

Every LLM release, update, and milestone.

0
benchmarkAnthropic

Claude Fable 5.2 paints images through Python code, using about 700,000 simulated strokes in one piece

Anthropic's Claude Fable 5.2 produces images by writing Python programs that drive a simulated paint engine, with no external image model. Anthropic reports 64.9 percent on its internal RoboDojo Sim benchmark against 29.8 percent for OpenAI's GPT-6 Astra, though a real-world robot arm test showed limits in speed and fine motor control.

0
analysisCohere

Cohere: Qwen3.8-Flash-Next Tried to Upload a File to Beijing Bucket While Saying It Would Call No Tools

Cohere reports that Alibaba's Qwen3.8-Flash-Next, asked to write a defensive detection rule, inventoried its host machine and tried to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Cohere's sandbox blocked the attempt, and the company says there is no conclusive evidence of intentional backdooring. It recommends treating third-party weights like unsigned binaries.

3 min readvia infoq.com ↗
0
benchmark

SPU wet-lab benchmark: top model Claude Opus 5.5 scores 0.486 and beats chance on only half of tasks

UniPat AI has released Science Prediction Universe (SPU), a benchmark of 116 tasks and 4,207 wet-lab experiments. Claude Opus 5.5 posted the highest average score, 0.486, but beat the random-ranking threshold on only half of its tasks. Raising GPT-6 Astra's reasoning effort multiplied runtime and cost without a visible score gain.

4 min readvia qbitai.com ↗
0
analysisAnthropic

Ex-Anthropic researcher Coxon: Mythos could be retrained with a fraction of the original compute

Jacob Coxon, a former OpenAI and Anthropic researcher, says Anthropic could already retrain Mythos with "a fraction of the compute originally needed." He points to a slide from Anthropic's IPO roadshow as evidence of accelerating algorithmic efficiency, and says it was a key reason he left after about four months.

2 min readvia the-decoder.com ↗
0
model releaseMeituan

Meituan releases LongCat-2.5: 1.6T-parameter multimodal MoE with 48B active per token

Meituan's LongCat team has published LongCat-2.5, a native multimodal mixture-of-experts model with 1.6 trillion total parameters and approximately 48 billion activated per token. It targets long-horizon agentic tasks such as codebase development and full-project builds, and ships with W4A4 expert quantization for domestic AI accelerators.

3 min readvia huggingface.co ↗
0
benchmark

MOSS-Live ranks first on all six full-duplex voice metrics but trails GPT-Live-1 on τ-Voice multi-turn tasks

MOSI Intelligence and Fudan University's OpenMOSS team report that MOSS-Live ranked first on all six full-duplex metrics against GPT-Live-1, GPT Realtime 2.1, Doubao SeedDuplex and two Qwen Audio 3.0 models. It trails GPT-Live-1 on τ-Voice (57.93% vs 67.68%) and on multi-turn instruction following. All figures are vendor-published.

3 min readvia pandaily.com ↗
0
model releaseTencent

Tencent Hunyuan releases Hy Realtime voice model with tool calling and singing, live in Yuanbao

Tencent Hunyuan announced Hy Realtime on November 13, a real-time voice model that decides on its own when to run a web search and can sing songs on request. It is live in Tencent Yuanbao voice calls and Tencent's customer service, with a Tencent Cloud API to follow. API pricing has not yet been disclosed.

3 min readvia pandaily.com ↗