Qwen3.8 Max
By Alibaba · released 2026-08-03 (5 weeks) · snapshot 2026-09-09
- Intelligence index
- 40.30
- Cost to run the AA index
- $3172
- Price per 1M tokens (in / out)
- $2.00 / $6.00
- End-to-end time
- 65.8s
Is Qwen3.8 Max on the Pareto frontier, and what beats it?
Off the frontier
No. GPT-6 Astra (high) is at least as smart for the same money or less.
Next step up: GPT-6 Astra (xhigh), smarter but dearer.
| # | Model | Creator | Intelligence | Total cost | E2E total | Open | Released | Frontier |
|---|---|---|---|---|---|---|---|---|
| 1 | Qwen3.8 Max | Alibaba | 40.30 | $3172 | 65.8s | 2026-08-03 | ||
| 2 | GPT-6 Astra (high) | OpenAI | 51.00 | $2917 | 55.6s | 2026-09-03 | Frontier | |
| 3 | GPT-6 Astra (xhigh) | OpenAI | 52.50 | $3803 | 178.2s | 2026-09-03 | Frontier |
Open in the picker Daily check URL: 200 on the frontier, 409 once beaten
Qwen3.8 Max today
Intelligence: Artificial Analysis Intelligence Index. Cost: total cost to run that benchmark, so reasoning tokens count. API id: qwen3-8-max.
Compare head to head
Suggested: what beats it, its effort levels, and frontier models near its price.
- Qwen3.8 Max vs GPT-6 Astra (high)
- Qwen3.8 Max vs GPT-6 Astra (xhigh)
- Qwen3.8 Max vs GPT-5.6 Sol (high)
- Qwen3.8 Max vs GPT-6 Astra (low)
- Qwen3.8 Max vs Muse Spark 1.3 (max)
- Qwen3.8 Max vs GPT-6 Astra (medium)
- Qwen3.8 Max vs GPT-6 Astra (max)
- Qwen3.8 Max vs Claude Fable 5.1 (Adaptive Reasoning, Xhigh Effort, Default Fallback)
What the Qwen3.8 Max release did to the frontier on 2026-08-03
None of its 1 effort levels reached the frontier at launch.
Closest miss: Qwen3.8 Max, 2.2 points below GPT-5.6 Sol (high) at 2.1× the cost.
Should you use Qwen3.8 Max?
Should you use Qwen3.8 Max?
Only when it fits your setup: GPT-6 Astra (high) is at least as smart for the same money or less, so Qwen3.8 Max is off the value frontier today.
What is better and cheaper than Qwen3.8 Max?
GPT-6 Astra (high) — at least as smart for the same money or less.
How much does Qwen3.8 Max cost?
$2.00/$6.00 in/out per 1M tokens, and $3172 to run the full Artificial Analysis Intelligence Index, the honest all-in number for reasoning models.
Is Qwen3.8 Max on the Pareto frontier?
No: GPT-6 Astra (high) beats it on intelligence, cost, or both.
Frontier membership by axis pair
Other axis pairs where Qwen3.8 Max is Pareto-optimal.
| Y axis | X axis | On the frontier |
|---|---|---|
| AA intelligence index | AA total cost to run (USD) | |
| AA intelligence index | End-to-end total seconds | |
| AA intelligence index | Output speed tokens/sec |
Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/models/qwen3-8-max · Markdown: /api/v1/models/qwen3-8-max?format=md · llms.txt · agent skill