Qwen 3.7 confirms the multi-vendor open-frontier stabilization — six vendors shipping at near-frontier-lab cadence
Pre-2026 open-source had a recurring single-vendor pull-ahead pattern: Llama 2 dominated, then Mistral, then Llama 3, then briefly DeepSeek V3. H1 2026 instead shows six vendors shipping at comparable cadence with comparable capability. The category structure now matches the closed-source frontier-lab landscape in durability.
Qwen 3.7's release continuing the Qwen 3.x line is unremarkable as a single event — but as part of the broader six-vendor pattern, it's structurally important. The six vendors (Qwen, DeepSeek, Llama, Kimi, Mistral, GLM) all ship near-frontier capability at comparable cadence. No vendor pulls dramatically ahead for more than a release cycle; no vendor falls dramatically behind.
What stabilization means for procurement
Open-source vendor selection now optimizes on capability-shape fit rather than on 'which open model is best right now.' The six-vendor landscape covers different workload specializations: Qwen for multilingual (201 languages), DeepSeek for cost-optimized cluster deployment, Llama 4 Scout for ultra-long-context, Kimi for agentic workloads, Mistral for efficiency, GLM for Chinese-language deployment. Selection criteria match workload requirements rather than vendor reputation.
The release-rhythm change
Pre-2026 open-source releases were monolithic — one major version, six-to-twelve month gaps. The Qwen 3.x cadence (3.5, 3.6, 3.7 within H1 2026) compresses the rhythm to match closed-source frontier labs. The implication for procurement is that open-source model selection now requires the same version-tracking discipline that closed-source vendor management requires.
What stays uncertain
The six-vendor stability could fracture if any vendor ships a step-change capability advance that the others can't match within 6-9 months. The historical pattern (Llama 2, Mistral, Llama 3, DeepSeek V3) suggests the field doesn't sustain single-vendor leads for long, but past performance isn't a guarantee. The H2 2026 procurement strategy should commit to open-source for current workloads while monitoring for capability-gap inflection points that would justify reconsideration.
Codersera — Open-Source LLMs Landscape: Qwen, Llama, DeepSeek, Kimi (May 2026) → · LLM Stats — LLM News Today (June 2026) – AI Model Releases → · Featherless — Best Open-Source LLMs in 2026 →