bug-bounty249
google212
facebook172
microsoft169
apple126
rce97
exploit89
web352
open-source44
smart-contract42
writeup42
defi41
sqli39
aws38
ethereum38
dos36
docker36
ai-agents36
access-control35
cloudflare35
malware34
cve34
ssrf33
react32
xss31
account-takeover28
subdomain-takeover27
supply-chain26
oauth25
idor25
bragging-post24
smart-contract-vulnerability23
cors22
wordpress22
node22
browser22
privilege-escalation21
race-condition20
automation20
auth-bypass19
cloud19
pentest19
tool19
authentication-bypass18
machine-learning18
denial-of-service17
llm17
vulnerability-disclosure17
ctf17
rust16
0
2/10
Mixedbread releases Wholembed v3, a multimodal multilingual retrieval model that achieves state-of-the-art performance on LIMIT and BrowseComp-Plus benchmarks, outperforming existing semantic search models and becoming the first semantic model to surpass lexical-based retrieval on structured-text-heavy documents.
retrieval
embeddings
semantic-search
multimodal
multilingual
information-retrieval
ai-model
benchmark
late-interaction
Mixedbread
Wholembed v3
LIMIT
BrowseComp-Plus
Cohere Embed 4
OpenAI Text Embedding 3 Large
Voyage 4 Large
Gemini Embedding 2
BM25
0
1/10
Cumulus Labs launches IonRouter, a low-cost inference API optimized for open-source and fine-tuned models, backed by IonAttention—a custom C++ inference runtime designed specifically for NVIDIA GH200 hardware architecture that achieves 588 tokens/s on multimodal workloads through novel optimizations around cache coherence, KV block writeback, and attention scheduling.
inference-api
gpu-optimization
ml-infrastructure
llm
cuda
gpu-orchestration
hardware-specific-optimization
multimodal
startup
IonRouter
Cumulus Labs
IonAttention
TensorDock
Palantir
Together AI
Fireworks
Modal
RunPod
vLLM
GH200
OpenAI
Veer
Suryaa