Fugu Max vs Kimi K3
Side-by-side comparison of API pricing, specs, benchmarks, and capabilities
| Specification |
by Sakana AI
|
by Moonshot
|
|---|---|---|
| Overview | ||
| Status | Current mid tier | Current flagship Open weights |
| Released | ||
| Pricing per million tokens | ||
| Input | $2.00/Mtok | $3.00/Mtok |
| Output | $6.00/Mtok | $15.00/Mtok |
| Blended avg | $3.00/Mtok | $6.00/Mtok |
| Cached input | $0.250/Mtok | $0.300/Mtok |
| Price basis | Sakana AI Token Plan (pay-as-you-go) list price, per 1M tokens, flat at any context length — the card states 'Fixed rates regardless of context length', unlike the context-tiered Fugu Ultra. web_search and web_fetch tool calls are billed separately at $0.007 per call and are not part of the per-token rate. The monthly subscription tiers (Standard $20, Pro $100, Max $200) bundle usage allowances rather than per-token rates and are not the list price. Sakana states the service is not offered to users in EU or EEA member states. | — |
| Specifications | ||
| Context window | 1M tokens | 1M tokens |
| Parameters | Proprietary | Proprietary |
| Speed (TPS) | — | — |
| Modalities | ||
| Input |
textimage
|
textimage
|
| Benchmarks sources: n/a / Vellum LLM Leaderboard | ||
| Avg benchmark score | — | 83.3 |
| Perf / dollar | — | 13.9 |
| GPQA Diamond | — | 93.5 |
| SWE-Bench Verified | — | 93.4 |
| Terminal-Bench 2.1 | — | 88.3 |
| MCP Atlas | — | 84.2 |
| BrowseComp | — | 91.2 |
| Humanity's Last Exam | — | 56 |
| HumanEval | — | 76.8 |
| Independent composite scores each on its own scale — not part of the average above | ||
| AA Intelligence Index | — | 43.8 |
| AA Agentic Index | — | 50.6 |
| AA-Omniscience Index (−100–100) | — | 19.7 |
| GDPval-AA v2 | — | 1583.5 ELO |
| LMSYS Chatbot Arena | — | 1485.3 ELO |
| Providers | ||
| Available from |
Sakana AI — $2.00/$6.00/Mtok
|
Moonshot — $3.00/$15.00/Mtok
|
Cost at scale
1M tokens · 50/50 input/output| Volume | Fugu Max | Kimi K3 | Savings |
|---|---|---|---|
| 1M tokens | $3 | $6 | $3 (50%) |
| 10M tokens | $30 | $60 | $30 (50%) |
| 100M tokens | $300 | $600 | $300 (50%) |
| 1000M tokens | $3000 | $6000 | $3000 (50%) |
When to pick which
distilled from the pricing and spec data above- cost dominates: $3.00/Mtok blended vs $6.00 — 50% less on the same 50/50 token mix
- your workload re-reads context (agents, RAG, long chats): cached input costs $0.300/Mtok — 90% off its list input price
- you want open weights — self-host it, fine-tune it, or exit the API entirely; Fugu Max is closed
Summary
Fugu Max by Sakana AI costs $2.00/Mtok input and $6.00/Mtok output, with a 1M-token context window. It supports text, image input.
Kimi K3 by Moonshot costs $3.00/Mtok input and $15.00/Mtok output, with a 1M-token context window. It supports text, image input.
On a blended cost basis, Fugu Max is 50% cheaper than Kimi K3.
The two aren't directly comparable on average benchmark score: Kimi K3 has published per-benchmark results, while Fugu Max does not yet.
Note: Pricing is per million tokens. Actual costs vary with usage patterns, prompt caching, and batch discounts. Always verify against official provider pricing pages.