# Tagged reasoning-model

> Announcements and real-world deployments of AI, checked against their sources and scored: open-weights releases, decision models in the wild, and what is actually in use.

- [Reflection releases Beam, a 501B open-weights MoE under Apache 2.0](https://shipped.anth.us/open-weights/reflection-beam-1-0-501b-a23b/index.md): Beam is Reflection's first open model: 501B total and 23B active parameters, 1M-token context, under Apache 2.0. (2026-10-07)
- [Aleph Alpha releases Kolibri 1, a 78B German and English reasoning MoE under Apache 2.0](https://shipped.anth.us/open-weights/aleph-alpha-kolibri-1/index.md): Kolibri 1 is a 78B mixture-of-experts reasoning model with 3.46B active parameters, tuned for German and English, under Apache 2.0. (2026-10-03)
- [Xiaomi MiMo releases MiMo-V2.6-Pro-RL, a 1.02T omnimodal MoE under MIT](https://shipped.anth.us/open-weights/xiaomimimo-mimo-v2-6-pro-rl/index.md): MiMo-V2.6-Pro-RL is the flagship of the MiMo-V2.6 series: 1.02T total, 42B active, 1M context, text, image, video and audio in. (2026-09-21)
- [JetBrains releases Mellum2.1 Thinking, a 12B coding MoE with 2.5B active, under Apache 2.0](https://shipped.anth.us/open-weights/jetbrains-mellum2-1-12b-a2-5b-thinking/index.md): Mellum2.1 Thinking is a 12B mixture-of-experts reasoning model with 2.5B active, retrained mostly with reinforcement learning on real repositories. (2026-09-20)
- [DeepSeek releases V4.1-Flash, a multimodal MoE with a much smaller KV cache, under MIT](https://shipped.anth.us/open-weights/deepseek-ai-deepseek-v4-1-flash/index.md): DeepSeek-V4.1-Flash is a 552B-backbone multimodal MoE with 1M-token context and a 890-byte-per-token KV cache, under MIT. (2026-09-10)
- [Z.ai releases GLM-5.3, a 753B open-weights model aimed at coding and cyber tasks](https://shipped.anth.us/open-weights/zai-org-glm-5-3/index.md): GLM-5.3 reuses the GLM-5.2 base with new post-training. The weights are public under a custom glm-5.3 licence. (2026-08-25)
- [Mistral releases Medium 3.5, a dense 128B model with a modified MIT licence](https://shipped.anth.us/open-weights/mistralai-mistral-medium-3-5-128b/index.md): Mistral Medium 3.5 is a dense 128B instruct, reasoning and coding model with 256k context. The licence excludes large-revenue companies. (2026-03-31)
- [NVIDIA releases Nemotron 3 Nano 30B-A3B, a hybrid Mamba-MoE reasoning model](https://shipped.anth.us/open-weights/nvidia-nemotron-3-nano-30b-a3b/index.md): Nemotron 3 Nano is a 30B hybrid Mamba-2 and MoE model with 3.5B active parameters, under the NVIDIA Nemotron Open Model License. (2025-12-04)
- [Ai2 releases Olmo 3, with a fully documented 32B Think model under Apache 2.0](https://shipped.anth.us/open-weights/allenai-olmo-3-32b-think/index.md): Olmo 3 has 7B and 32B Instruct and Think variants. Ai2 releases code, checkpoints and training details alongside the weights. (2025-11-19)
- [Moonshot releases Kimi K2 Thinking, a 1T MoE thinking agent with native INT4](https://shipped.anth.us/open-weights/moonshotai-kimi-k2-thinking/index.md): Kimi K2 Thinking is a 1T-parameter MoE with 32B active, 256K context and native INT4 quantisation, under a modified MIT licence. (2025-11-04)