REVA screens new functions for vulnerabilities in about one second
On PrimeVul, REVA with GLM-5.3-Flash matches VulTriage accuracy while running over 250 times faster.
46
- Mentions · 7d2
- Mentions · 30d10
- All mentions18
- First seenAug 27, 2026
- Last mentioned4d ago
Timeline
18 articles · newest firstCoding-agent harnesses cost more without better benchmark scores
Harness swaps changed outcomes about as often as reruns, while system prompts made some tasks cost up to three times more.
46
HABIT-Bench finds agent memory weak at inferring user habits
The best of eight memory methods scored 33.61% macro accuracy on Qwen3-235B-A22B and GLM-5.3-Flash across four domains.
45
MiniMax releases M3.1 Flash, with open weights due October 23
API pricing is $0.15 per million input tokens and $0.60 per million output tokens, with cached input at $0.015.
52
GLM-Legal: Advancing Agentic Legal Intelligence
Z.ai unveiled GLM-Legal, a GLM-5.3-Flash legal agent that matches frontier models on Harvey, LegalBench, and PLawBench.
52
Import AI 476: Creativity and RSI; worse-than-China open weight models; AI goes to space
Humanity’s Sixth Sense finds GPT-6 Astra at 53.6 pass@1 versus a 93.1 human baseline on intuitive visual reasoning.
52
REMORY: Learning Residual Memory for Context Compaction
REMORY adds residual soft memory tokens so agents compact context without losing later decision quality.
48
Write Like It's 1866: LLMs Relearn Telegraphese
A benchmark finds LLMs can store records in telegraph-style cablese, cutting tokens by up to 49 percent without hurting later answers.
46
autotrust/GLM5.3-Flash-E224-DGX-Spark — new model trending #29 on Hugging Face
autotrust released an unofficial pruned NVFP4 GLM-5.3-Flash build that fits two NVIDIA DGX Sparks.
44
Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks
Cantina and Yeta Labs released apex-flash-1, an open model that solved 40 of 60 bug tasks.
71
Anthropic says Zhipu's open-weight GLM-5.3 nearly matches Claude Mythos Preview at building exploits
Anthropic says open-weight GLM-5.3 nearly matches Claude Mythos Preview at building software exploits.
78
The current balance of power in open models
A report finds Chinese open-weight AI models now outperform American counterparts in downloads and benchmarks.
45
How Fragile Is Safety Alignment at Frontier Scale? A Single-Direction Attack on a 320B MoE
Researchers show directional ablation breaks refusal in GLM-5.3-Flash, a 320B-parameter MoE, cutting refusal by 41–89 points across seven benchmarks.
45
Latest open artifacts (#24): Motif-3, GLM-5.3, Hy4-preview and open model licenses
Interconnects surveys new open models—Motif-3, GLM-5.3, Hy4-preview—while analyzing a licensing split: Western labs opening up, Chinese frontier labs getting restrictive.
52
Sparks Fly: NVIDIA Accelerates Local AI at IFA 2026
NVIDIA announces local AI push at IFA 2026 with faster llama.cpp/vLLM inference, PAIR routing tool, and October RTX Spark PCs.
55
[AINews] Fal’s H3 Max Live breaks the infinite videogen barrier
Fal post-trained MiniMax H3 into a 'Max' variant with 35x-faster inference, enabling faster-than-realtime AI video generation and infinite streams.
50
[AINews] OpenAI to reach AGI bar by end-2026
OpenAI chief scientist Jakub Pachocki says unreleased Astra model meets the 'Automated AI Research Intern' goal; Altman expects internal AGI declaration by December 2026.
48
[AINews] NVIDIA buys HuggingFace for $13B, as OpenAI publishes their HF incident retro
Z.ai released open-weight GLM-5.3-Flash (320B/18B active, 1M context, MIT) while Nvidia confirmed buying Hugging Face for $13B.
78
Appears with
- NVIDIAvendor · 4
- Z.aivendor · 4
- Kimi K3ai model · 4
- Z.aiorganization · 4
- OpenAIorganization · 3
- Nvidiaorganization · 3
- Googleorganization · 3
- Qwen3.8-Flash-Nextai model · 3
- GLM-5.3ai model · 3
- DeepSeekvendor · 3
- Metavendor · 3
- Googlevendor · 3
- qwen3.8-27bai model · 3
- Alibabaorganization · 2
- Hy4-previewai model · 2
- DGX Sparkproduct · 2
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.