There’s still room to beat Deepseek in communicating clearly and following instructions to the letter (it’s very over-eager imo). I can put it in subagent but for main model, as far as open models go, there’s still no alternative to GLM-5.3 Flash.

(Though to be frank even Sol&Astra have the same problems, and Deepseek is probably heavily "inspired". I guess it may come down to preference though the new Terra 6.1 does very favorably on both of those aspects. I really do prefer to work with it. Wish there was some kind of Terra Super with the same benchmarks as the bigger models.)

> yet still cannot even beat MiniMax M3.1 Flash (220B) despite having 2.7x more parameters

MiniMax M3.1 Flash is 428B total parameters, so the size ratio is more like 1.4x. Model card here: https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash

The number HF calculates from the safetensors metadata is often wrong (e.g. counting u8-packed 4-bit weights as half a weight each) so you need to take the number from the actual model card.

I wouldn't bet on the numbers of compute reported by Chinese labs. Those Singapore Nvidia orders had to go somewhere.

Yes, Mira Murati's Thinking Machines. They are apparently trying to raise $5B at $40B, after raising $2B at $10B last year.

In their last technical report (for the original Inkling model), they mentioned 'To bootstrap post-training, we ran an initial SFT on synthetic data generated by open-weights models including Kimi K2.5. The bootstrap accounts for a small fraction of compute, with the majority being employed for large-scale RL on synthetic and human-created environments.' and in their most recent one they said 'The H3 and H4 mixtures consisted mostly of trajectories sampled from earlier models including Kimi K3 and our own models, filtered for correctness and alignment.' So it does seem like traces from larger models do contribute. I think Microsoft is one of the few companies that has actively made an effort to avoid distillation with their MAI series.

Everyone is building on everything that comes before.

Houses are not built from scratch either. The know-how, power, water, roads, materials…