Alibaba moved Qwen3.8-Max out of preview and into general availability on August 3, 2026, two weeks after previewing the model at WAIC Shanghai with a single slide and a claim that it trailed only one rival. This time the flagship shipped with what the preview lacked: a published benchmark table, real per-token pricing, and a live production API. It’s a 2.4 trillion-parameter Mixture-of-Experts (MoE) model with roughly 95 billion active parameters per token, a 1 million-token context window, and native text, image, and video input. The open-weight release Alibaba promised for the week of August 10, however, has not appeared on Hugging Face or ModelScope as of August 11.
What Alibaba shipped with Qwen3.8-Max
The general availability release is built on the same architecture Qwen3.7-Max introduced, scaled up and extended to multimodal input for the first time above the trillion-parameter mark. Key specifications include:
- 2.4 trillion total parameters, sparse MoE, roughly 95 billion active per token
- 1,000,000-token context window (991K max input, 131K max output, up to 262K for reasoning chains)
- Native text, image, and video input with text output
- Input price $2.00 per million tokens, output price $6.00 per million tokens
- Cached input from $0.25/M (implicit) to $0.17/M (explicit read)
- Rate limits of 2M tokens per minute and 15,000 requests per minute
- OpenAI-compatible and Anthropic Messages-compatible API endpoints, live day one
- Model ID qwen3.8-max on QwenCloud, Alibaba Cloud Model Studio, QwenWork, and the Vercel AI Gateway at zero markup
Alibaba is framing the release less as a chatbot upgrade and more as an autonomous-agent flagship.
How does Qwen3.8-Max perform on benchmarks?
Alibaba’s published benchmark table compares Qwen3.8-Max against three current flagships. Every number is Alibaba’s own vendor-run evaluation, not an independent measurement.
| Benchmark | Qwen3.8-Max | Claude Fable 5 | GPT-5.6 | Claude Opus 4.8 |
|---|---|---|---|---|
| OSWorld-Verified | 86.1 | ~85.0 | 83.2 | Not published |
| PaperBench | 93.0 | 88.8 | 90.5 | 80.3 |
| Terminal-Bench 2.1 | 86.6 | 84.6 | 88.8 | 84.6 |
| SWE-bench Pro | 67.7 | 80.0 | 64.6 | 69.2 |
| GPQA Diamond | 92.6 | 92.6 | 94.1 | 92.0 |
| IFBench | 82.8 | 63.5 | 72.7 | 62.2 |
| HLE (Humanity’s Last Exam) | 43.6 | 53.3 | 47.2 | 45.7 |
The pattern is directionally clear: Qwen3.8-Max wins on agentic computer-use and long-document tasks (OSWorld-Verified, PaperBench, IFBench) and holds its own on general terminal agentic work. It loses to Claude Fable 5 by 12 points on SWE-bench Pro, the harder professional coding benchmark, and finishes last of the four flagships on HLE broad-knowledge reasoning, nearly 10 points behind Fable 5.
What can Qwen3.8-Max actually do?
Three capability demonstrations anchor the launch.
Long-horizon autonomous coding
The headline demo is oh-my-cli, a command-line agent framework the model built and continues to run on its own. It turns incoming requests into GitHub issues, claims them through a state machine, writes code, runs end-to-end tests, and merges its own pull requests without human intervention. As of July 30 the project had produced 265 commits, 127 pull requests, and 151 issues over 16 days. The repository has continued adding commits since then and remained actively updated around the time of the GA launch, with hundreds of additional commits and an Apache-2.0 license.
Research and competition tasks
Alibaba reports Qwen3.8-Max reproduced a published paper on data selection for LLM reasoning, writing roughly 7,600 lines of code and running 33 GPU training rounds across five days to land a +2.71 point improvement on AIME24 over the original paper’s method. In a separate 24-hour coding competition, the model’s entry beat 458 of 526 human teams, placing in the 87th percentile.
Native multimodal agents at scale
Qwen3.8-Max processes documents past 200 pages and video past 100 hours using what Alibaba calls video memory graphs, pairing GUI screen operation with visual feedback loops for self-verification, measured internally against Alibaba’s RecreationBench.
Pricing and availability
Qwen3.8-Max is live now on QwenCloud at $2.00 per million input tokens and $6.00 per million output tokens, with cached input as low as $0.17/M on explicit reads. That undercuts Kimi K3’s $3.00/$15.00 pricing by a wide margin and roughly matches Qwen3.7-Max’s prior $2.50/$7.50 rate card.
Access runs through the QwenCloud API, Alibaba Cloud Model Studio (international scope), and the Vercel AI Gateway at zero markup under the slug alibaba/qwen3.8-max. The day-one Anthropic Messages-compatible endpoint means Claude Code can point at Qwen3.8-Max by changing ANTHROPIC_BASE_URL and ANTHROPIC_MODEL, with no other workflow changes, the cheapest path for a side-by-side comparison against Claude models in an existing agent setup.
Where is the open-weight release?
The open-weight release is the unresolved piece. Alibaba said weights for Qwen3.8-Max and a smaller Qwen3.8-27B would land on Hugging Face and ModelScope during the week of August 10. As of August 11 no repository has appeared for either model and no license has been named. Whether a Max-class open weight would ship under the permissive Apache-2.0 license used for smaller Qwen releases such as Qwen3.6-27B, or under something more restrictive, remains an open question.
Strengths and weaknesses
Strengths
- Real, checkable benchmark table replaces the preview’s unverified second-only-to-Fable-5 marketing line
- Leads the four-way comparison on OSWorld-Verified, PaperBench, and IFBench
- $2.00/$6.00 pricing undercuts Kimi K3 by a wide margin while adding native multimodal input
- Day-one Anthropic-compatible endpoint makes it a drop-in swap for Claude Code and similar agent systems
- The oh-my-cli autonomous coding project is a live, ongoing, publicly auditable demonstration rather than a one-time benchmark run
Weaknesses
- Trails Claude Fable 5 by 12 points on SWE-bench Pro, the harder of the two coding benchmarks in Alibaba’s table
- Finishes last of four flagships on HLE, nearly 10 points behind Fable 5
- Every benchmark number is vendor-run by Alibaba, with no independent evaluator yet to copy the full table
- Promised open weights for the week of August 10 have not shipped as of August 11, with no license confirmed
- The active-parameter figure (~95B) comes from third-party reporting rather than an Alibaba-published technical report or model card
FAQ
Is Qwen3.8-Max open source?
Not yet. Alibaba said open weights for Qwen3.8-Max and a smaller Qwen3.8-27B would land on Hugging Face and ModelScope during the week of August 10, 2026, but as of August 11 no repository has appeared and no license has been confirmed.
How much does Qwen3.8-Max cost?
$2.00 per million input tokens and $6.00 per million output tokens on QwenCloud, with cached reads as low as $0.17 per million tokens. That price is roughly one-third of Kimi K3’s per-token cost.
Does Qwen3.8-Max beat GPT-5.6 and Claude Fable 5?
It depends on the task. Alibaba’s published table shows Qwen3.8-Max ahead on OSWorld-Verified (86.1), PaperBench (93.0), and IFBench (82.8), but trailing Claude Fable 5 by 12 points on SWE-bench Pro (67.7 vs 80.0) and finishing behind all three rivals on HLE (43.6).
How many parameters does Qwen3.8-Max have active per token?
Third-party reporting puts it at roughly 95 billion active parameters out of 2.4 trillion total. Alibaba has not published an official technical report confirming this figure.
Can I use Qwen3.8-Max with Claude Code?
Yes. Alibaba shipped an Anthropic Messages-compatible endpoint on day one. Pointing ANTHROPIC_BASE_URL at the DashScope endpoint and setting ANTHROPIC_MODEL to qwen3.8-max runs existing Claude Code workflows against it with no further changes.
Related coverage
- Alibaba unveils Qwen3.8-Max, a 2.4 trillion parameter AI model
- Zhipu AI Releases GLM-5.3, Claims Strongest Open-Weights Coding Model
Want help with this in Myrtle Beach or Long Island? Talk to us.
NewSunSEO
Grow in Google and AI Search
AI-SEO, local search, paid ads and websites from one team, with offices on Long Island and in Myrtle Beach.
- AI-SEO
- Local AI-SEO
- AI Consulting
- PPC
- Web Design
- Web Development
- Long Island Office(631) 320-1700
- Myrtle Beach Office(843) 492-7131
Request a Free Proposal
Tell us about your business and where you want to grow.
We will reply by email to talk through your goals and what a proposal would include.
We use your details only to reply to you and send what you asked for. See our Privacy Policy.
