New ModelsOpen weights
DeepSWE-Preview
DeepSWE-Preview hits 59% SWE-Bench Verified with pure RL on Qwen3-32B
Agentica and collaborators (with guest Michael Luo of UC Berkeley) released DeepSWE-Preview, a fully open-sourced RL-trained coding agent built on Qwen3-32B that reached 59% on SWE-Bench Verified, a top open result in a benchmark dominated by closed systems. The team published training methodology and weights, emphasizing reproducible reward design and verification over sealed benchmark numbers.
59% SWE-Bench Verified
New Models
Qwen-TTS
Alibaba launches Qwen-TTS with human-level bilingual naturalness
The Qwen team released Qwen-TTS, a bilingual Chinese/English text-to-speech model claiming human-level naturalness, available via API with a Hugging Face demo space. It was the second voice release of the week alongside Kyutai TTS.
New ModelsOpen weights
ERNIE 4.5
Baidu open-sources ERNIE 4.5, a 10-model multimodal family
Baidu open-sourced the ERNIE 4.5 series, a family of 10 models ranging from 424B down to 0.3B parameters with multimodal capabilities, reportedly beating o1 on DocVQA. The release marks a sharp reversal from Baidu's previous anti-open-source posture and another sign that Chinese labs are setting the pace in open source.
10 ERNIE 4.5 models
New Models
Chai-2
Chai Discovery's Chai-2 enables zero-shot antibody design
Chai Discovery introduced Chai-2, a model for zero-shot antibody design that generates candidate antibodies without iterative lab screening. Mentioned in the show notes tools section as one of the week's notable science releases.
New ModelsOpen weights
Pangu Pro MoE
Huawei's Pangu Pro MoE: 72B model trained entirely on Ascend NPUs
Huawei released Pangu Pro, a 72B-parameter MoE trained on its own Ascend NPUs rather than Nvidia or AMD hardware, hitting 1,528 tokens/sec and pretrained on 13T tokens. The panel framed it as the geopolitical open-model story of the week, showing how far Chinese compute stacks have advanced under sanctions.
New ModelsOpen weights
Kyutai TTS
Kyutai releases open low-latency TTS for English and French
Kyutai Labs released an open 1.6B-parameter text-to-speech model with low latency and high voice similarity in English and French. It was one of two TTS launches closing out the episode, underscoring how quickly multimodal product quality is rising.
New Models
Cypher Alpha
Mystery 1M-context model 'Cypher Alpha' appears free on OpenRouter
A stealth model called Cypher Alpha showed up on OpenRouter with a free 1M-token context window, with the panel speculating it could be Amazon Titan. Alex used it as an example of how model releases increasingly arrive as anonymous market probes rather than tidy launches.
New ModelsOpen weights
Hunyuan-A13B-Instruct
Tencent ships Hunyuan-A13B: 80B MoE with only 13B active params
Tencent released Hunyuan-A13B-Instruct, an 80B-parameter MoE that activates only 13B parameters at inference while keeping a 256K context window. Built by the team with WizardLM lineage, it posts strong reasoning benchmarks and feels unusually practical for its class, though the panel flagged its license limits.
13B Hunyuan active params
