{"slug":"moonshot-ai-s-kimi-k3-launch","label":"Qwen 3.8","item_count":3,"day_count":3,"source_count":3,"first_seen":"2026-07-22T14:09:04+00:00","last_updated":"2026-08-09T01:35:57+00:00","generated_at":"2026-08-12T10:12:25.915148+00:00","sources":["interconnects","latent_space","search_cn_open_weight_labs"],"days":[{"date":"2026-07-22","items":[{"title":"Open models recap: more on Kimi K3, Qwen 3.8, Xi's WAIC speech, distillation, the open-closed gap, and what's next","url":"https://www.interconnects.ai/p/open-models-recap-more-on-kimi-k3","source":"interconnects","type":"news","summary_1line":"A podcast with Florian Brand.","sid":"9347a64bc25f44cc","published":"2026-07-22T14:09:04+00:00","editor_note":"First item to pair Kimi K3 and Qwen 3.8 as rivals within the wider open-vs-closed model debate."}]},{"date":"2026-08-04","items":[{"title":"[AINews] Qwen 3.8 Max(2.4T) and 27B, new open weights models for Coding and Cowork","url":"https://www.latent.space/p/ainews-qwen-38-max24t-and-27b-new","source":"latent_space","type":"news","summary_1line":"Qwen is so back!","sid":"c09b9b5f698dc233","published":"2026-08-04T03:49:14+00:00","editor_note":"Alibaba adds a 2.4T Max tier and a 27B coding/agent-focused variant, Qwen 3.8's first concrete expansion in this thread."}]},{"date":"2026-08-09","items":[{"title":"Kimi K3 vs Qwen 3.8: Which Open Weight Chinese AI is Better? - Memeburn","url":"https://news.google.com/rss/articles/CBMiU0FVX3lxTE1ZVFhaZy15aUNMRmNEMnR2LXE0WmgzVG9sLXF0TWh6d05hY3lhQy01UnllZ0xiT1RHaE1YazlteVFNYnRHNFN2SENULWZDcTFKdVJJ?oc=5","source":"search_cn_open_weight_labs","type":"news","summary_1line":"Kimi K3 vs Qwen 3.8: Which Open Weight Chinese AI is Better? Memeburn","sid":"1f245454c88a0663","published":"2026-08-09T01:35:57+00:00","editor_note":"Memeburn's head-to-head is the thread's first direct comparison verdict between the two models."}]}],"editorial":{"tldr":"A July 22 podcast recap first grouped Kimi K3 and Qwen 3.8 as rival open-weight entries in China's open-vs-closed AI debate. On August 4, Alibaba expanded Qwen 3.8 with a 2.4-trillion-parameter Max tier and a 27B variant aimed at coding and collaborative-agent use, positioning it as a direct Kimi K3 competitor.","stale":false,"whats_new":"Memeburn published a head-to-head comparison of Kimi K3 and Qwen 3.8 on August 9, the thread's first direct model-vs-model verdict.","why_it_matters":"Platform engineers evaluating open-weight Chinese models now have two actively-updated families to weigh — Kimi K3 and Qwen 3.8's new 27B coding/agent tier — rather than a single default choice.","take_for_builders":"Before picking either model for a coding or agent workload, run Qwen 3.8's 27B tier against Kimi K3 on your own eval set rather than taking Memeburn's comparison or either vendor's benchmark claims at face value.","beats":[{"kicker":"CONTEXT","tone":"neutral","headline":"Podcast recap frames Kimi K3 and Qwen 3.8 as rival open-weight entries","summary":"Interconnects' recap ties the two models to the broader open-vs-closed model debate around Xi's WAIC speech and distillation.","sids":["9347a64bc25f44cc"]},{"kicker":"QWEN 3.8 EXPANDS","tone":"rising","headline":"Alibaba ships Qwen 3.8 Max (2.4T) and a 27B coding/agent tier","summary":"The new 27B variant specifically targets coding and collaborative-agent use, widening Qwen 3.8's competitive footprint against Kimi K3.","sids":["c09b9b5f698dc233"]},{"kicker":"NOW","tone":"now","headline":"Memeburn runs a direct Kimi K3 vs Qwen 3.8 comparison","summary":"The thread's first head-to-head verdict between the two open-weight Chinese model families.","sids":["1f245454c88a0663"]}],"open_questions":["Does Memeburn's head-to-head rely on independent benchmarks or vendor-published numbers?","Does Qwen 3.8's 27B coding/agent tier hold up against Kimi K3 on real tool-use and coding workloads, not just headline comparisons?"],"generated_at":"2026-08-09T15:30:00+00:00"}}