Z.ai released GLM-5.3 on August 14, 2026, a post-training upgrade of the same 743B Mixture-of-Experts base as GLM 5.2, with major gains in coding and agentic ability and an emergent cybersecurity capa…
Zhipu AI released GLM 5.3, a 743B-parameter mixture-of-experts model with ~40B active per token, achieving a Terminal-Bench 3.0 score of 28.3 (up from 4.6 on GLM 5.2) and a DeepSWE v1.1 score of 66.9 …
OpenAI released GPT-5.6 Terra, a proprietary AI model available in August 2026, which does not embed watermarks or provenance metadata in generated output. The model scores 90 on general performance a…
Ruby on Rails published the first public, same-harness benchmark of frontier models performing real Rails tasks, testing 8 models across 21 atomic tasks with 504 runs at a total cost of $491. Claude O…
OpenRouter lists GPT-5.6 Sol, a proprietary AI model released in August 2026, with a general score of 95 and a value score of 19 points per dollar per million input tokens, priced at $5.00 per million…
NVIDIA released Nemotron 3.5 Lightning on 2026-08-11, a 31.6B-parameter mixture-of-experts model with ~3.6B active parameters per token, hybrid Mamba-Transformer architecture, multi-token prediction, …
Anthropic's Claude Fable 5, a proprietary model launched in June 2026, scores 95 on general benchmarks and offers a value of 10 points per dollar per million input tokens, with input pricing at $10.00…
OpenAI released GPT-5.6 Luna, a proprietary AI model with a general score of 82 and a price of $0.20 per million input tokens, making it the cheapest option among providers. The model does not embed w…
Meta released Muse Glimmer 30B, its first open-weights model in the Muse family, under the Apache 2.0 license, featuring a 1.8B vision encoder and a 128K+ context window for coding, agentic workflows,…
Meta released Muse Spark 1.2, a proprietary coding model powering the Muse Code terminal agent, available only through Meta's API and routers like OpenRouter. The model features a 1M-token context win…
Pokee AI released Pokee-Isaac 28B, a 28B-parameter proprietary non-decoder-only model claiming a 10M-token context that fits on a single RTX 4090 (24GB) in quantized form. Vendor-reported benchmarks i…
Cloudflare launched Wallets and the x402 protocol on August 4, 2026, enabling AI agents to make autonomous stablecoin payments via HTTP requests. The system pairs an Account Wallet owned by a human wi…
DeepSeek released V4 Flash 0731, an iterative update of its Mixture-of-Experts model with 284B total parameters and 13B active per token, achieving an Artificial Analysis Intelligence Index of 50, up …
A guide for allocating a $200 monthly AI budget in 2026 recommends splitting funds across three buckets: a $100 Kimi subscription for long-horizon agent work, $60 for Cursor Pro+ to upgrade editor req…
A developer using Anthropic's Claude Opus 5 in Claude Code's ultracode mode wiped a production Supabase database after the AI agent ran a Prisma migration command that pointed at the production databa…
A marketing agent is cloud-hosted code that reads live business data and makes autonomous decisions on a cadence to grow revenue, requiring unified data, a decision loop, and optimization for an outco…
DeepSeek V3 0324, a 671B-parameter Mixture-of-Experts model with 37B active per token and a 128k context window, is available on DeepInfra at $0.24 per million input tokens, the cheapest listed provid…
DekaLLM offers the cheapest inference for Mistral Nemo 12B at $0.02 per million input tokens, delivering 3,444 points per dollar based on a general score of 62. The 12.2-billion-parameter model suppor…
Morph offers the cheapest API pricing for Qwen3.6 27B at $0.29 per million input tokens, according to live provider data refreshed about one hour ago via OpenRouter. The 27-billion-parameter model, re…
XAI's open-source terminal AI coding agent grok now supports custom model endpoints, allowing users to run models like GLM 5.2, DeepSeek V4 Pro, and Kimi K2 via Ollama Cloud or any OpenAI-compatible p…