NVIDIA releases Nemotron 3 Super, a 120B-parameter open model with only 12B active at inference, combining Mamba-2 and Transformer layers for agentic AI workloads with a 1M token context window.
vLLM v0.17.0 adds FlashAttention 4, elastic expert parallelism for live MoE rescaling, full Qwen3.5 support, and a performance-mode flag, all in 699 commits from 272 contributors.
Kimi K2.5 leads every coding benchmark, but Qwen3.5-35B-A3B delivers 87-93% of that performance at 3-4x lower cost and runs on a single consumer GPU. Here is the full breakdown.
Zhipu AI's GLM-5 is a 744B MoE model with 40B active parameters, trained on 100K Huawei Ascend chips, scoring 77.8% SWE-bench and 50 on Artificial Analysis Intelligence Index - MIT licensed.
Zhipu AI's 744B open-source model GLM-5 was trained entirely on Huawei Ascend chips and now competes with GPT-5.2 and Claude Opus on major benchmarks.
Two Chinese open-weight trillion-parameter MoE models with ~32B active parameters each - DeepSeek V4 bets on cost and context, Kimi K2.5 bets on Agent Swarm and verified benchmarks.
A pre-release comparison of DeepSeek V3.2 and V4 - examining the generational leap from 671B text-only to a trillion-parameter natively multimodal model with 1M context.
DeepSeek V4 is an unreleased trillion-parameter MoE model with ~32B active parameters, native multimodal capabilities, a 1M-token context window, and optimization for Huawei Ascend chips - expected in the first week of March 2026.
Moonshot AI's Kimi K2.5 is a 1T-parameter MoE model activating 32B per token with native multimodal vision via MoonViT-3D, Agent Swarm coordination of up to 100 sub-agents via PARL, and top-tier math and coding benchmarks under a modified MIT license.
Head-to-head comparison of Moonshot AI's Kimi K2.5 and Anthropic's Claude Opus 4.6 - an open-weight MoE powerhouse against the reigning agentic coding champion.
A direct comparison of Kimi K2.5 and DeepSeek V3.2 - two open-weight Chinese MoE models fighting for different corners of the cost-performance frontier.
Comparing Kimi K2.5 and Gemini 2.5 Flash-Lite - Moonshot AI's 1T parameter open-weight powerhouse against Google's cheapest and fastest inference option.
