https://modelsatwork.news/article/apertus-1-5-open-weights-8b-70b-262k-context-apache-2-full-training-data Models at Worken 2026-10-15T14:10:00.000Z Apertus 1.5: open 8B and 70B models, 262k context, Apache 2.0 Model release, Open-weight models, Open source, Evals, Governance https://modelsatwork.news/article/jevcrawl-open-source-agent-crawler-0-025-per-task-vs-0-194-llm-link-picking Models at Worken 2026-10-15T13:05:00.000Z Jevcrawl crawler for agents: $0.025 a task, 40% correct Open source, Agents, Agent cost, Show HN, Benchmarks https://modelsatwork.news/article/aws-sagemaker-jumpstart-model-test-script-one-prompt-qwen3-5-27b-9-seconds-delete-endpoint Models at Worken 2026-10-15T12:10:00.000Z AWS shows a script to test SageMaker JumpStart models AWS, Amazon SageMaker AI, Evals, Latency, Vendor-published https://modelsatwork.news/article/hugging-face-tokenizers-bert-50x-faster-ascii-skip-embedding-rag-pipelines Models at Worken 2026-10-15T10:05:00.000Z Hugging Face makes BERT tokenizing up to 50x faster Inference, RAG, Cost, Open source, Vector search https://modelsatwork.news/article/aws-agentcore-powerpoint-generator-sample-0-04-dollars-per-deck-vendor-published Models at Worken 2026-10-15T01:57:00.000Z AWS AgentCore deck generator sample: ~$0.04 and 70–90 seconds Agents, Amazon Bedrock AgentCore, Vendor-published, Cost, RAG https://modelsatwork.news/article/graphicle-open-source-rust-code-graph-memory-for-coding-agents-kubernetes-3-min-vs-26-min Models at Worken 2026-10-15T01:30:00.000Z Graphicle: open-source code graph and memory for coding agents, runs locally Coding agents, Open source, Agents, Cost, Community https://modelsatwork.news/article/anthropic-midterms-safeguards-multi-step-misuse-classifiers-0-002-percent-violations Models at Worken 2026-10-14T22:00:00.000Z Anthropic adds classifiers for multi-step election misuse before US midterms Anthropic, Governance, Safeguards, Evals, Agent safety https://modelsatwork.news/article/microsoft-thinkingbox-agents-graded-on-database-state-47-percent-pass-all-20-runs Models at Worken 2026-10-14T19:59:00.000Z Microsoft's ThinkingBox: best agents pass all 20 runs on 47.5% of tasks Agent evaluation, Evals, Benchmarks, Cost, Human in the loop https://modelsatwork.news/article/parsimony-open-benchmark-counts-how-much-code-ai-coding-models-add Models at Worken 2026-10-14T18:58:00.000Z Parsimony open benchmark ranks AI coding models by code added Coding agents, Evals, Open source, Benchmarks, Claude https://modelsatwork.news/article/anthropic-claude-api-opentelemetry-export-beta-traces-managed-agents Models at Worken 2026-10-14T17:58:00.000Z Claude API adds OpenTelemetry trace export in beta, off by default Claude, Anthropic, Monitoring, Managed Agents, Data retention https://modelsatwork.news/article/aws-agentic-idp-agentcore-sample-58-dollars-month-synthetic-data-vendor-published Models at Worken 2026-10-14T16:56:00.000Z AWS agentic invoice sample costs ~$0.055 a document; accuracy untested Agents, Multi-agent, Amazon Bedrock AgentCore, Vendor-published, Cost https://modelsatwork.news/article/lighton-colgrep-agent-2b-claude-code-search-34-percent-fewer-tokens-latency-up Models at Worken 2026-10-14T16:01:00.000Z LightOn's ColGrep-Agent cut Claude Code search tokens 34%, but slower Coding agents, Token cost, Open source, Claude, Inference https://modelsatwork.news/article/timescale-pg-vectorsearch-prism-index-alpha-postgres-18-pgvector-compatible Models at Worken 2026-10-14T15:05:00.000Z Timescale's pg_vectorsearch 0.1.0 adds a PRISM index to Postgres; it is alpha RAG, Open source, Vector search, Postgres, Community https://modelsatwork.news/article/tamnoon-tami-bedrock-cloud-remediation-90-percent-exposure-zero-production-issues-vendor-published Models at Worken 2026-10-14T13:58:00.000Z Tamnoon's Tami agent fixes cloud findings on Bedrock; 90% claim is vendor's Agents, Security, Amazon Bedrock, Vendor-published, Human in the loop https://modelsatwork.news/article/aws-sagemaker-moe-kernel-tuning-10-percent-throughput-9-percent-ttft-vllm Models at Worken 2026-10-14T13:00:00.000Z AWS: MoE kernel fixes give up to 10% more throughput on H200 vLLM, Mixture-of-experts, Inference, Amazon SageMaker AI, Cost, AWS https://modelsatwork.news/article/aws-bedrock-managed-entitlements-private-pricing-grants-stay-disabled-until-activated Models at Worken 2026-10-14T11:05:00.000Z Bedrock private pricing stays off until a grant is activated Amazon Bedrock, AWS Marketplace, AWS License Manager, Cost governance, Private pricing, AWS https://modelsatwork.news/article/google-play-private-ai-studio-apps-no-developer-account-oct-2026 Models at Worken 2026-10-14T10:00:00.000Z Google Play lets AI Studio apps run privately, no developer account Google Play, Google AI Studio, Vibe coding, Android, Shadow IT, Internal tools https://modelsatwork.news/article/grok-4-7-reasoning-effort-bedrock-72-runs-low-effort-one-eighth-tokens Models at Worken 2026-10-13T21:57:00.000Z AWS test: Grok 4.7 at low effort got 15 of 18 right on 1/8 the tokens Grok 4.7, Amazon Bedrock, Reasoning effort, Agent evaluation, Token cost, AWS https://modelsatwork.news/article/anchor-browser-agentcore-payments-x402-session-flag-real-payments-default-off Models at Worken 2026-10-13T21:00:00.000Z Anchor Browser adds x402 payments for agents via AgentCore Anchor Browser, Amazon Bedrock AgentCore, x402, Agent payments, Browser agents, AWS https://modelsatwork.news/article/google-gemini-4-argon-generally-available-promo-price-ends-30-november Models at Worken 2026-10-13T18:58:00.000Z Google's Gemini 4 Argon is GA; $2/$10 price rises to $4/$20 on 1 Dec Google, Gemini 4 Argon, Gemini API, Pricing, Release, Model selection https://modelsatwork.news/article/anthropic-scion-source-control-30000-merges-a-day-seven-engineers Models at Worken 2026-10-13T17:57:00.000Z Anthropic replaced git with Scion as merges hit 30,000 a day Anthropic, Source control, Scion, Coding agents, CI, Monorepo, Claude Tag https://modelsatwork.news/article/orca-security-discovery-ai-bedrock-query-cost-down-90-percent-tokens-2m-to-140k Models at Worken 2026-10-13T16:58:00.000Z Orca Security cut its AI query agent's cost 90% on Amazon Bedrock Orca Security, AWS, Amazon Bedrock, RAG, Reranking, Model routing, Evals, Claude Haiku 4.5 https://modelsatwork.news/article/shipstation-deckhand-slack-it-help-desk-agent-812-of-1221-chats-no-ticket Models at Worken 2026-10-13T15:58:00.000Z ShipStation's Slack IT agent kept 812 of 1,221 chats out of Jira ShipStation, AWS, Amazon Bedrock AgentCore, IT help desk, Slack, Ticket deflection, Strands Agents https://modelsatwork.news/article/aws-sagemaker-llama-cpp-decision-model-gpu-4-9x-faster-p95-32x-accuracy-77-percent Models at Worken 2026-10-13T14:58:00.000Z AWS: a 4B decision model on SageMaker scored 77.5% on CPU, 77.9% on GPU AWS, SageMaker, llama.cpp, Decision models, Routing, Latency, Inference https://modelsatwork.news/article/apple-minions-idle-mac-gpus-inference-queue-hours-to-minutes-no-benchmarks Models at Worken 2026-10-13T14:15:00.000Z Apple's Minions pools idle Macs for inference; no benchmarks yet Apple, Minions, Inference, GPU capacity, Ollama, Research https://modelsatwork.news/briefing/2026-10-15 Models at Worken 2026-10-15T10:30:00.000Z The briefing, 2026-10-15: Microsoft and Hugging Face say the best agent passed all 20 runs on only 47.5% of 507 workflows https://modelsatwork.news/briefing/2026-10-14 Models at Worken 2026-10-14T10:30:00.000Z The briefing, 2026-10-14: Orca Security says a nine-stage pipeline cut its Bedrock agent's cost 90% and latency from 60 to 15 seconds