Category desk

Research

Technical deep dives, reproducible tests, and tool evaluations.

ThreatFrontier data poster: GPT-6 lineup with Astra at $10/$50, Sol at $2/$10 and Luna at $0.10/$0.50 per million tokens, a 50% price cut from GPT-5.6 Sol, Sol DeepSWE 68.8%, 1.05M-token context, released Sep 3 and Sep 22 2026.

GPT-6 Launches With Near-Frontier Intelligence at Half the Price

GPT-6 Sol and Luna halve OpenAI's API prices yet score near rival flagships on coding, while flagship Astra is OpenAI's first Critical-rated cyber model.

Hub-and-spoke data poster titled Qwen 4 Confirmed In Training, with satellites for Qwen 4.5, Qwen 5, the Zhenwu V900 accelerator and Alibaba Cloud capacity targets.

Qwen 4 Research

Alibaba Confirms Qwen 4 in Training as Apsara Roadmap Targets 10T Parameters

Alibaba confirmed Qwen 4 is in training at Apsara 2026, with Qwen 4.5 and Qwen 5 targeting 5-10T parameters and the Zhenwu V900 chip due Q1 2027.

Data poster headed CYBERGYM CEILING BROKEN for Xiaomi MiMo-V2.6, with a stat rail showing CyberGym Pro 94.0, CyberGym Flash 95.1, ExploitBench Pro 47.9, licence MIT, context 1M, beside a node diagram linking MiMo-V2.6-Pro 1.02T, MiMo-V2.6-Flash 309B, open weights, no API gate and 7000 RL environments.

Mimo V2 6 Research

Xiaomi's MiMo-V2.6 Breaks the CyberGym Ceiling and Ships MIT Weights the Same Day

Xiaomi's MiMo-V2.6 clears CyberGym at 94.0 under an MIT licence, but ExploitBench at 47.9 shows the audit-to-exploit gap is still wide open.

Vllm Vs Sglang Research

vLLM vs. SGLang in 2026: Benchmarking RadixAttention vs. PagedAttention for Local Agents & RAG

In agentic workflows, prefill latency and prefix caching matter more than raw tok/s. We benchmark SGLang RadixAttention vs vLLM PagedAttention and APC.

Apple Silicon Research

Mac Studio (M4/M5 Ultra 128GB–256GB) vs. Dual RTX 5090 (64GB): The $6,000 Local LLM Workstation Shootout

Choosing a $6,000 local AI workstation: Apple Silicon Unified Memory (128GB-256GB) vs Dual RTX 5090 (64GB). Sizing 70B/236B MoE models, TTFT, and TCO.

Local Llm Inference Research

The RTX 5090 (32GB) Local LLM Deployment Blueprint: NVFP4 vs. FP8, SGLang vs. vLLM, and Context Limits

Can the RTX 5090 run 70B models? We break down 32GB VRAM sizing, Blackwell NVFP4 vs FP8 benchmarks, KV cache math, and vLLM vs SGLang configurations.

Claude Code Research

Claude Code Pricing & Token Economics: Subscription vs. Pay-As-You-Go API

Claude Code CLI burns 400k to 2.5M tokens per task. We analyze subscription tiers (Pro, Max 5x, Max 20x) vs direct API pricing with prompt caching.

Chatgpt Pro Research

Cursor Ultra vs. Claude Max 20x vs. ChatGPT Pro: The $200/Month Developer Tier Showdown

Cursor Ultra, Claude Max 20x, and ChatGPT Pro all charge $200/month. We stress-test real rate limits, usage pools, and token ROI during 8-hour sprint days.