Blog

OpenAI expands the GPT-6 universe with Sol and Luna, delivering massive performance gains in coding benchmarks while slashing API costs by 50%.

Anthropic's Claude Opus 5.5 arrives as a milestone release, delivering Fable-level performance at a 40% lower cost with unprecedented coding and alignment capabilities.

With a massive jump in coding agent benchmarks and a specialized RL training regimen, xAI's Grok 4.7 is setting a new standard for long-context engineering tasks.

Xiaomi shatters the open-source ceiling with MiMo-V2.6, a 1T-parameter sparse MoE model that rivals GPT-5.6 and Claude Opus 5 in agentic reasoning.

PrismML expands its Bonsai series with 2B to 27B ternary-quantized models, enabling high-reasoning performance on consumer hardware via Qwen3.8 lineage.

DeepSeek's V4.1 Flash breaks the efficiency barrier, outperforming flagship models in agentic coding benchmarks while slashing operational costs.

Nex AGI has released Nex-N2.5, a powerhouse family of open-weights models designed to bridge the gap between LLMs and autonomous digital agents.

IFM has released K2 Horizon, the largest and most transparent open-source model fleet in AI history, spanning from 0.9B to 375B parameters.

OpenAI has officially released GPT-6 Astra, a milestone model featuring a 1M+ context window and granular reasoning control, signaling the arrival of the AGI era.