Large language model coverage: specifications, context windows, API pricing and benchmark results for frontier and open-weight models, updated as they ship.
Chinese outlet 36 Kr reports that Anthropic has officially released Fable 5.1, presenting it as a demonstration of next-generation capabilities. The report is headline-level, so here is what is actually confirmed, what remains open, and how the Fable tier compares on price and context.
xAI has reportedly released Grok 4.6, according to CoinGape. The update marks the latest step in the Grok model family, known in part for its distinctive grok spicy mode setting.
OpenAI has published a post saying two configuration settings tripled its ARC-AGI-3 scores, with the-decoder.com reporting the company now claims GPT-5.6 Sol beats Opus 5 on the benchmark when run through its latest API.
A 2.8T-parameter open MoE scoring 57 on the AA Intelligence Index – above Claude Opus 4.8. The specs, the benchmarks, the new $3/$15 pricing, and whether you can run it.
SpaceXAI’s newly released Grok 4.5 dramatically undercuts Anthropic and OpenAI on price, according to SiliconANGLE, intensifying a frontier-model price war that developers have watched build for months.
The full Ollama models list for 2026 — every popular model with parameters, download size and RAM/VRAM needs, plus how to list and pull models.
NVIDIA’s Nemotron 3 Nano Omni packs vision, audio, video, and text into one open 30B-A3B model with only 3B active parameters. Here’s what it does, how good it is, and what it takes to run.
Moonshot’s Kimi K2.7 Code is a 1T-parameter open-weight model built only for agentic coding. We verify the specs, the vendor benchmarks, the int4 hardware reality, and whether its aggressive $0.95/$4.00 pricing actually beats the closed frontier.
As of June 19, 2026, GPT-5.6 has no model card, benchmarks, or pricing from OpenAI. We separate the confirmed record from the loud leaks (1.5M context, Mini/Pro variants, a ‘meaningful improvement’ quote) and show how a hypothetical GPT-5.6 would land against Claude Fable 5, Gemini 3.5, and the open-weight wave from China.
Zhipu’s GLM 5.2 pairs a 1-million-token context with MIT-licensed open weights and a drop-in Anthropic-compatible API. Here are the confirmed specs, the benchmarks that landed with the weights, real access and self-hosting costs, and who should actually use it.
Ollama is the headless runtime developers wire into everything; Jan is the open-source desktop app that puts a ChatGPT-style UI and MCP tools in front of local models. Here’s how they actually stack up in mid-2026.
A practical, zero-to-running guide to LM Studio in 2026: setup on every OS, choosing models and quantization, the built-in OpenAI-compatible server, and the hardware you actually need.
