Monthly roundup
June 2026 in AI models
A release-wave summary for readers who want the important changes without pretending every launch is a reason to switch.
The signal
9 releases plus 1 material status update
Anthropic, OpenAI, Z.ai, Moonshot AI, Microsoft AI, Qwen account for 9 tracked release events. 7 catalogued events link to a detailed comparison profile. 2 launch guides are available for deeper trade-off coverage. There is no single winner because the constraints are different.
At a glance
- Release events
- 9
- Profile links
- 7
What shipped and changed
Limited previewGPT-5.6 Luna
OpenAI
Fast, lower-cost GPT-5.6 tier included in the initial limited preview.
Date note: General availability followed on July 9, 2026.
OpenAI preview announcement- Limited preview
GPT-5.6 Sol
OpenAI
Limited preview for selected trusted partners and organisations.
Date note: General availability followed on July 9, 2026.
OpenAI preview announcement - Limited preview
GPT-5.6 Terra
OpenAI
Balanced GPT-5.6 tier included in the initial limited preview.
Date note: General availability followed on July 9, 2026.
OpenAI preview announcement 
ReleaseGLM-5.2
Z.ai
Date note: Initial Coding Plan rollout preceded the June 16 public release post; QwenCloud records API availability on June 26.
Z.ai announcement- Status update
Claude Fable 5 access
Anthropic
Access was suspended for all users after a US government export-control directive.
Anthropic redeployment update 
ReleaseSecondary receiptKimi K2.7 Code
Moonshot AI
Date note: AI Release Tracker reports June 12; the official Kimi model page is dated July 22, so this earlier date remains a secondary-source launch event.
AI Release Tracker latest index (secondary source)
Limited previewMAI-Thinking-1
Microsoft AI
Date note: Private preview on Microsoft Foundry at announcement; not a generally available API release.
Microsoft AI announcement
Read the important launches
Anthropic
Claude Sonnet 5: mid-tier tokens that can hide agentic cost
Sonnet 5 looks cheaper per token than Opus 5, but long agentic runs can produce outsized output bills — the decision is completed-task cost, not the rate card alone.
Z.ai
GLM 5.2: open weights with a million-token context
GLM 5.2 pairs open weights with long context and mid-tier coding results — interesting for self-hosting teams that can carry integration and serving work.