token.app › Compare › Qwen3.8 2.4T A95B vs Mercury 3
Qwen3.8 2.4T A95B vs Mercury 3
Mercury 3 is 18× cheaper per token on a blended 3:1 basis. Qwen3.8 2.4T A95B accepts the larger context window (1.0M).
Head to head
Blended price uses a 3:1 input:output mix
| Qwen3.8 2.4T A95B | Mercury 3 | |
|---|---|---|
| Provider | Alibaba | Inception |
| Input $/1M | $2.00 | $0.100 |
| Output $/1M | $6.00 | $0.350 |
| Blended $/1M | $3.00 | $0.163 |
| Context window | 1.0M | 1M |
| Max output | 909K | 131K |
| Released | Aug 2026 | Oct 2026 |
| Licence | Open weights | Proprietary |
Choose Qwen3.8 2.4T A95B if…
- Larger context window (1.0M)
- Open weights — self-hostable
Choose Mercury 3 if…
- Costs less per token — $0.163 vs $3.00 blended
- Newer model (released Oct 2026)
FAQ
Which is better, Qwen3.8 2.4T A95B or Mercury 3?
We do not hold independently-run benchmark scores covering both Qwen3.8 2.4T A95B and Mercury 3, so we make no quality claim. Their pricing and specifications are compared on this page.
Is Qwen3.8 2.4T A95B cheaper than Mercury 3?
Mercury 3 is cheaper: $0.163 versus $3.00 per million tokens blended at 3:1 input:output. Input alone: $2.00 vs $0.100. Output alone: $6.00 vs $0.350.
What context windows do Qwen3.8 2.4T A95B and Mercury 3 have?
Qwen3.8 2.4T A95B: 1,048,576 tokens. Mercury 3: 1,000,000 tokens.
Who makes Qwen3.8 2.4T A95B and Mercury 3?
Qwen3.8 2.4T A95B is made by Alibaba. Mercury 3 is made by Inception.
Related comparisons