Is Qwen3.5 397B A17B a good deal?
Whether anything in this catalogue beats Qwen3.5 397B A17B on both quality and price, and what you give up if it does. A computation on the current catalogue, not an opinion.
12 undominated of 144 · Oct 16, 2026
As of Oct 16, 2026, MiMo-V2.6-Pro scores higher and costs less than Qwen3.5 397B A17B for Balanced on LMArena, but it is not a drop-in. You would give up: max output.
Inspect model evidence Compare differences & requirements
MiMo-V2.6-Pro scores 52.4 points higher and costs 38% less, but you would give up: max output.
LMArena Elo · higher is better
Scale starts at 1040 Elo
1437.9
Effective $/M · Balanced · lower is better
$0.878/M
LMArena Elo · higher is better
Scale starts at 1040 Elo
1490.3
Effective $/M · Balanced · lower is better
$0.540/M
Qwen3.5 397B A17B takes text, image, video, returns up to 235,929 tokens from a 262,144-token context, and is listed by 10 sellers.
Compared against 144 rated, priced models on this lens: nothing dominates it outright, 5 more dominate it but give something up. Qwen3.5 397B A17B scores 1437.9 at $0.878 per million tokens for this mix.
Higher score, lower price, named losses
Not a drop-in. The loss is why these are not a recommendation.
| Model | LMArena | Effective $/M | You give up |
|---|---|---|---|
| MiMo-V2.6-Pro | 1490.3 | $0.540/M | max output |
| GLM 5.3 Flash | 1471.0 | $0.238/MPrice under confirmation since Oct 12, 2026 | max output |
| MiMo-V2.5-Pro | 1465.1 | $0.544/MPrice under confirmation since Oct 16, 2026 | max output, image, video |
| DeepSeek V4.1 Flash | 1462.5 | $0.182/MPrice under confirmation since Oct 12, 2026 | video |
| Kimi K2.6 | 1455.4 | $0.825/M | video |
The link keeps the model and mix, using the current catalogue. Monthly spend and switching cost stay in this browser tab and are left out of the link.
Saved decision references
Save the model, workload, catalogue date and capability-preservation rule in this browser. Spend, switching cost and bill contents are not saved. No account or notifications.
Constraint: replacements must preserve the model’s capabilities; any losses remain named trade-offs.
Historical decisions cannot be fully replayed from saved references: past prices, scores and capability evidence are not stored.