$ diff fugu-max longcat-2-0
Fugu Max vs LongCat-2.0
Values come from Fugu Max and LongCat-2.0, where each is cited to its source. This page states no benchmark result and ranks neither model — it puts two published specifications next to each other. Where a lab has not published a figure, the row says so rather than guessing.
Sakana AIFugu Max
- context
- 1M
- weights
- —
- $/M in
- $2
- $/M out
- $6
- context
- 1M
- weights
- open
- $/M in
- $0.3
- $/M out
- $1.2
What actually differs
- Context window
- Identical — both accept 1M tokens, so context length is not a reason to pick either.
- Input price
- LongCat-2.0 at $0.3/M against $2/M — 6.7× cheaper to feed. Standard rates: one of these labs also quotes a lower cached-input tier, which applies only when a prefix is reused — see the full spec below.
- Output price
- LongCat-2.0 at $1.2/M against $6/M — 5.0× cheaper to generate. Output dominates the bill on most agentic workloads, where the model writes far more than it reads.
- Recency
- Fugu Max shipped 73 days after LongCat-2.0 (2026-09-11 vs 2026-06-30).
Full spec
| Attribute | Fugu Max | LongCat-2.0 |
|---|---|---|
| Developer | Sakana AI | Meituan |
| Released | 2026-09-11 | 2026-06-30 |
| Context window | 1,000,000 | 1 million tokens (LongCat Sparse Attention / LSA) |
| Pricing | $2/M input · $6/M output | $0.30/M input · $1.20/M output ($0.006/M cached input) — OpenRouter catalogue, read 2026-07-28. Open weights, so this is hosted-inference pricing rather than a licence fee. |
| License | not recorded | MIT (commercially permissive) |
| Availability | Sakana API (OpenAI-compatible), OpenRouter, NanoGPT, Kilo Gateway, LLM Gateway | Open-sourced 2026-06-30; previously served on OpenRouter under the codename "Owl Alpha" |