Alibaba Qwen

Tier A · Strong with clear operating rules.

Qwen3.8-Max

A low-A specialist with a genuine eye for website design.

Canonical model record

Current identity, limits, and pricing

Provider source · checked 2026-08-14 ↗

Status: Current
API model ID: qwen3.8-max
Context: 1M
Max output: 131K
API price / 1M tokens: $2 input · $6 output

Visual prompt runs

Benchmark runs

Open each generated scene, or compare the same prompt across models.

6 runs

City Scroll Journey

Cinematic scroll journey testing AI-generated scene continuity, scroll-scrubbed camera motion, and art-directed landing craft.

Open run →Compare this model →

Ember Glider

Sunset gliding journey testing flight energy management, checkpoint flow, and atmospheric scene craft.

Open run →Compare this model →

Low-Poly Tower Defense

Diorama tower defense testing economy balance, wave design, placement rules, and combat readability.

Open run →Compare this model →

Neon Drift

Synthwave time-trial racing testing drift physics, lap timing, ghost replay, and unlock progression.

Open run →Compare this model →

Starfall Arena

Neon arena survival testing wave escalation, upgrade builds, particle feedback, and boss design.

Open run →Compare this model →

Stormwind Trebuchet Simulator

Counterweight siege simulation testing coupled mechanics, trajectory prediction, projectile cameras, and interactive tuning.

Open run →Compare this model →

Superbash commentary

Our take

Back to the full tier list →

Qwen3.8-Max earns low A tier after a strong first-pass website benchmark. The team’s clearest reason to choose it is front-end design: it produced a polished animated landing page and correctly made use of the available image-generation skill without needing to be walked through every step.

Best for

Website design
Front-end and visual coding
Animated landing pages
Tool-aware design workflows

Watch out

Low A rather than an all-purpose frontier default
The showcased run used an external image-generation integration
Premium Chinese models are no longer dramatically cheaper than US peers

Why it is ranked here

Six published Superbash visual runs provide direct output evidence for the current model label.
The team was impressed by its first-try website design and its ability to invoke the right available skill.
It sits at the bottom of A tier because its standout case is specific rather than universal.

Evidence and commentary

2026-08-17

Superbash editorial model ranking

Takeaway: Qwen3.8-Max is currently placed in Tier A.

The August 2026 editorial roster places Qwen3.8-Max at rank 8.

Open source →

2026-08-14

Qwen3.8-Max model page

Takeaway: Qwen3.8-Max is the verified current flagship identity behind this catalog entry.

The provider page confirms the final model ID, multimodal inputs, one-million-token context, and current API prices.

Open source →

Keep learning

Official benchmark profile

Qwen3.8-Max is current, but its numeric benchmark table is not public.

The benchmark folders retain their original run provenance. The current catalog now resolves them to Qwen3.8-Max, while this score panel stays empty until Qwen publishes a source-specific numeric table.

Qwen sourceAugust 2026Official model page →