🎁 New Sign up free, 10 calls on us. Up to $1, no card needed.

Qwen3.8 Max Released 2026-08-03

Alibaba chatVisionCodeReasoningTool callingLong contextStructured outputStreamingPrompt caching
Input$2/M
Output$6/M
Cache read$0.25/M
Context984K
vs GPT-4o~60% cheaper

Provider list prices: no platform markup, pay-as-you-go. These are official list prices. Logged-in customers may see effective prices including workspace discounts on /console/pricing. Effective input at a 70% cache-hit rate:$0.775/M. Implicit context cache is automatic; an explicit cache_control mode offers deeper discounts (1,024-token minimum blocks, 5-minute TTL that resets on hit). Per-model cached-read prices are in the pricing table.

Use Qwen3.8 Max in 30 seconds

OpenAI-compatible: swap the base_url, keep your SDK. POST /v1/chat/completions

from openai import OpenAI

client = OpenAI(
    base_url="https://synthorai.io/v1",
    api_key="sk-syn-...",
)

resp = client.chat.completions.create(
    model="qwen3.8-max",
    messages=[{"role": "user", "content": "Summarize this diff"}],
    reasoning_effort="medium",
)
print(resp.choices[0].message.content)

Specs & limits

Tokens

Context window (vendor spec)
983,616
Max output (vendor spec)
131,072

Prompt caching

How it caches
automatic + explicit
Lifetime
explicit: 5m, reset on hit

Thinking

Parameter
reasoning_effort
Values
minimal · low · medium · high accepted set is the provider's

Model

Modalities
text + image → text
Parameters
2.4T total MoE
  • Flagship Qwen3.8 Max: natively vision-language, 2.4T-parameter MoE
  • thinking trace returned in reasoning_content
  • text and image in, text out

per Alibaba official docs ↗

About Qwen3.8 Max

2.4T-parameter MoE, natively vision-language
Text and image in, 131K max output
Tiered cache pricing for prompt-heavy loops

Qwen3.8 Max is Alibaba's flagship Qwen model, published on 3 August 2026.

  • Alibaba describes it as a natively vision-language Max model built on a Mixture-of-Experts architecture with 2.4 trillion parameters, and as the most capable model in the Qwen line to date.
  • It takes text and images and returns text, so a single request can mix a prompt with screenshots, charts or scanned pages instead of routing images through a separate vision model.
  • Thinking is part of the default behaviour: the model returns its reasoning trace in a separate reasoning_content field alongside the answer, which means short replies still consume reasoning tokens and a very small max_tokens budget can come back with an empty answer even though the request succeeded.
  • Function calling, strict JSON-schema structured output and streaming with usage in the final chunk are all available, so it drops into an existing OpenAI-compatible integration without special handling.
  • Context is just under one million tokens and a single response can run to 131,072 tokens, which is double the output ceiling of the previous Max generation and the concrete reason to move long-form generation onto it.
  • Pricing is tiered by cache behaviour rather than by context length: standard input, a cheaper rate for automatic cache hits, and a separate pair of rates for explicitly created and read cache entries, so prompt-heavy agent loops benefit materially from reusing a stable prefix.
  • Alibaba lists availability in its Beijing and Singapore regions.
  • Synthorai serves it through the OpenAI-compatible chat completions endpoint.

FAQ

Is the Qwen3.8 Max API free to try?

Yes: new accounts get 10 trial calls and up to $1 in free credit, no card required. At $2/M input tokens, that credit alone covers roughly 62 requests of ~8K tokens against Qwen3.8 Max.

What is Qwen3.8 Max best at?

2.4T-parameter MoE, natively vision-language; text and image in, 131K max output; tiered cache pricing for prompt-heavy loops. See the About section for the full picture from the vendor's own release notes.

How much does Qwen3.8 Max cost?

Qwen3.8 Max costs $2 per million input tokens and $6 per million output tokens on Synthorai. That is the provider's list price, with no platform markup. Cached input tokens bill at $0.25/M.

Does Qwen3.8 Max support prompt caching?

Yes: automatic caching is on by default, with an explicit mode for guaranteed savings. Cached input tokens bill at $0.25/M vs $2/M uncached (TTL explicit: 5m, reset on hit). Prompt caching guide →

How do I get access to Qwen3.8 Max?

Point your existing OpenAI SDK at base_url="https://synthorai.io/v1", set model="qwen3.8-max", and you're done. One API key covers every model on the gateway.

Related models

Get your free API key Compare your cost →