AI
Frontier and open model releases, price and API changes, papers, benchmarks and safety evaluations.

AI calorie estimates fail to scale with stated food weights
Multiplying each model’s calorie estimate per unit of weight by the stated food weight reduced errors for small and large portions.
Last 7 daysNov 17 – Nov 22
Thinking Machines Lab releases Inkling 2 Mini weights
Inkling 2 Mini runs its middle layers three times for every output token it generates.
AI illustrationAI model outputs first speech audio in 390 ms on NVIDIA GB10
A single model takes text, images and speech as inputs and uses a reference voice to synthesise speech.Simplex AI releases code for its on-device Linger AI agent
Adopting Linger depends on which tasks it can handle offline and whether the licence permits the intended use of the code.
AI illustrationLanguage model edits depend on evaluation conditions
Papers on refusal removal and vocabulary trimming assessed model behaviour after edits alongside the conditions used to evaluate it.
Runway releases robot-control model weights
Usage rights and links to video production are the criteria creators need to assess the model release.Small model decisions remain sensitive to who takes part
Pooling evaluations left participant effects intact, prompting researchers to recommend candidate sets and margins over forced rankings.
1 week agoNov 10 – Nov 15
Sophon tuning improves spatial reasoning in open models
Researchers traced spatial reasoning failures to a bottleneck in how models route information into answers.
Pakistan port AI design sets ownership and data boundaries
The proposal links operator ownership with domestic model execution and excludes sending customers’ cargo information abroad or to foreign model services.
Decider weight release carries repository access conditions
Decider weights are reportedly offered under an Apache licence, while instructions for using the full model include authentication for a private repository.
MiniMax unveils model for direct high-resolution images
DigitalToday reported that MiniMax's image model supports generation without separate upscaling, editing and distributed weights for developers to run it themselves.
Language gaps persist in document understanding benchmark
Researchers reported that larger models did not narrow performance gaps between languages in the MDU benchmark.
Real-time maths checks improve reported reasoning accuracy
A preprint reports higher symbolic maths accuracy alongside increased token use, while its abstract reports no independent replication.
NAVER Cloud's multimodal training throughput rises to 150.2%
A rebuilt vision encoder and rebalanced image load drove the gain, in a company-published result measured on one 30B model.
2 weeks agoNov 05 – Nov 09
llama-modchip fits 284B DeepSeek model into 64GB RAM, no GPU
The file shrinks to 56.56GiB, but unpatched llama.cpp, Ollama and LM Studio cannot run it.
Baidu opens ERNIE 5.5 Flash Base weights, but not for Korea
Baidu's pre-training-only base weights are free for commercial use up to a revenue threshold, but the licence does not cover the EU, the UK or South Korea.
Drex Max decision model scores 67.52 on Decision Index 0.4.1
The score and response times of the open model, which returns only a probability for each option, are the authors' own figures and have no independent replication.
Oct 2026Oct 25
Meta researchers release 120-task test for wearable AI aides
The benchmark scores the final state of a virtual world rather than written answers, but the repository notes list no pass rates by model.
18 in totalWritten by Gyeol HanRSS feed

