September 2026 AI Model Releases: GPT-6, Claude Opus 5.5, Gemini 3.8, DeepSeek, Qwen and GLM

By ChatGBot · September 29, 2026

September 2026 turned into one of the busiest months the AI world has seen, with more than 20 new models shipping in a two week stretch and a genuine price war breaking out. Here is what actually launched, what it costs, and why it matters if you use these models every day.

Claude Opus 5.5 leads the frontier

On September 22, Anthropic released Claude Opus 5.5, the first model in the Claude 5.5 family. It is priced at $4 per million input tokens and $20 per million output tokens, about 20% below Opus 5, with cached input dropping to $0.20. Anthropic says it beats the previous flagship on agentic benchmarks while making typical workloads roughly 40% cheaper to run. Sonnet 5.5 and Haiku 5.5 are expected in the following weeks. If you want the background, see our guide to Claude Opus and Claude models explained.

OpenAI answers with GPT-6 Sol and GPT-6 Luna

About ninety minutes after Anthropic, OpenAI shipped two new GPT-6 tiers. GPT-6 Sol ($2 / $10 per million tokens) targets heavier work like coding, while GPT-6 Luna ($0.10 / $0.50) is built for high volume tasks such as summarizing documents and extracting data. Luna sets a new mainstream price floor. This release was less about topping the leaderboard and more about cost per task. New here? Start with what is GPT-4o.

DeepSeek, Qwen and GLM keep the pressure on

• DeepSeek V4.1-Flash (Sep 10): $0.30 / $1.20 at peak and $0.15 / $0.60 off peak, with a 4x KV cache reduction for agent workloads. See is DeepSeek free.
• Qwen3.8 Max Prime (Alibaba, Sep 23): $4 / $12 per million tokens, sitting at the top of the Qwen line. More in what is Qwen.
• GLM-5.3-Flash (Z.ai): a list price of $0.15 / $0.50, one of the cheapest capable options around. Background in what is GLM AI.

Google ships Gemini 3.8 Flash

Google was busy as well. On September 2 it released Gemini 3.8 Flash, its third Flash model in six weeks, priced at $0.75 / $3.75 per million tokens through the end of 2026 before the rate doubles on January 1, 2027. It handles text, images, audio, video and PDF with a 1 million token context window, and is tuned for long horizon coding and autonomous agents. Google shipped it alongside a locked down security variant called 3.8 Flash Cyber, and says it beats its own 3.7 Flash across the board while edging out Claude Opus 5 on a few benchmarks. And there is more coming: Google has confirmed that Gemini 4, its next flagship, is already in post training and expected before the end of 2026, with a focus on coding and long horizon agents. More background in our Gemini chatbot guide.

What it means for you: a real price war

The headline of September 2026 is not a single winner, it is falling prices. With GPT-6 Luna at $0.10 per million tokens and strong options from DeepSeek, Qwen and GLM, the gap between premium and budget models is now about cost per task rather than raw capability. The smart move is no longer picking one model and hoping it fits every job. It is matching the model to the task: a frontier model like Claude Opus 5.5 or GPT-6 Sol for hard reasoning and coding, and a cheap, fast model for summaries and bulk work.

Use every new model in one place

You should not need five subscriptions and five logins to keep up with a month like this. ChatGBot gives you GPT, Claude, DeepSeek, Qwen and GLM in a single app, so you can switch models mid chat and always reach for the right one. That is the whole point: less juggling, one bill, every top model.

Frequently asked questions

What is the newest Claude model? Claude Opus 5.5, released September 22, 2026, is the newest and the first of the Claude 5.5 family, with Sonnet 5.5 and Haiku 5.5 expected soon.

What are GPT-6 Sol and Luna? They are two GPT-6 tiers from OpenAI. Sol is tuned for complex work like coding, and Luna is a low cost tier for high volume tasks.

Which new model is the cheapest? GPT-6 Luna, at $0.10 per million input tokens, set a new mainstream price floor in September 2026, with GLM and DeepSeek close behind.

Can I use all of these in one app? Yes. ChatGBot lets you use GPT, Claude, DeepSeek, Qwen and GLM together under one subscription.