Matthew Berman showed 5 AI tools in this video on . 1 is verified available in the EU. Each row below jumps to the exact second. Part of week 2026-32.
Tools shown in this video, with timestamps
New open-weights ~2.4T parameter model from Alibaba that is competitive with top US closed models on benchmarks (Terminal Bench 86.6 vs 84.6 for Fable), leads in multimodal, document and spatial categories. Highlighted for reproducing research papers with self-generated improvements and autonomous chip design. Priced at $2/M input and $6/M output on OpenRouter, far below GPT-5.6 Soul and Fable.
Closed-source US frontier model used as comparison; leads SWE-Bench Pro at 80 and visual agent/coding benchmarks. Priced at $10/M input and $50/M output, rumored at 7+ trillion parameters.
OpenAI model used as comparison, priced $5/M input and $30/M output; scored just above Qwen 3.8 Max on Terminal Bench. OpenAI said it helped improve efficiency enough to cut prices by 80%.
Prior Qwen model shown on Artificial Analysis chart at $1.28 per task completed, since the new Qwen 3.8 isn't yet tested there.
Moonshot's open-source model in the mid-2 trillion parameter range, shown on the Artificial Analysis cost-per-task chart at about 30-40% less than the most expensive Anthropic models.
