gpt-oss-20b vs Kimi K3
Pricing, context window and real answers (September 2026)
gpt-oss-20b (OpenAI) and Kimi K3 (Moonshot), compared as you can actually call them on FastMetal. Both are served from the same OpenAI-compatible endpoint and API key; switching is a change to the model string.
Specs and pricing
| Provider | OpenAI | Moonshot |
|---|---|---|
| Input (per 1M tokens) | ¥5.36 | ¥504 |
| Output (per 1M tokens) | ¥25.02 | ¥2,520 |
| Typical cost (1,000 in + 500 out tokens × 1,000 calls) | ¥18 | ¥1,764 |
| Context window | 131,072 tokens | 1,048,576 tokens |
| Release date | 8/5/2025 | 7/16/2026 |
| Input modalities | text | text, image |
| Throughput (p50) | — | — |
| Arena · overall | #252 · ELO 1,317 | #17 · ELO 1,485 |
| Arena · Japanese | #179 · ELO 1,256 | #2 · ELO 1,511 |
| Arena · coding | #232 · ELO 1,369 | #7 · ELO 1,538 |
Prices are FastMetal's yen rates before tax. Arena ranks are as of the latest leaderboard scrape.
Which should you pick?
- On price, gpt-oss-20b: ¥25.02 per 1M output tokens, 99% less than Kimi K3.
- On the Overall arena board, Kimi K3 ranks higher (#17; gpt-oss-20b is #252).
- On the Japanese arena board, Kimi K3 ranks higher (#2; gpt-oss-20b is #179).
- On the Coding arena board, Kimi K3 ranks higher (#7; gpt-oss-20b is #232).
- For long documents, Kimi K3: a 1,048,576-token context window against 131,072 for gpt-oss-20b.
- If in doubt, try both on the same key. Switching is a change to the model string, and each is billed at its own rate.
Real answers to the same prompts
Exactly what the FastMetal gateway returned, side by side. Not benchmark scores: actual output.
Count the number of 'r's in 'strawberry'
Count the number of 'r's in 'strawberry'. Explain your reasoning step by step.
gpt-oss-20b
Kimi K3
Debug This Error
I'm getting the following error in my Node.js application: TypeError: Cannot read properties of undefined (reading 'map') at UserList (/app/components/UserList.js:12:25) at renderWithHooks (/app/node_modules/rea…
gpt-oss-20b
Kimi K3
Code Review
Please review the following Python function and suggest improvements for readability, performance, and best practices: def get_data(url, retries=3): import requests import time for i in range(retries):…
gpt-oss-20b
Kimi K3
Frequently asked questions
- Which is cheaper, gpt-oss-20b or Kimi K3?
- Per 1M output tokens, gpt-oss-20b is ¥25.02 and Kimi K3 is ¥2,520 on FastMetal (yen, before tax), so gpt-oss-20b is cheaper.
- How do the context windows of gpt-oss-20b and Kimi K3 compare?
- gpt-oss-20b takes 131,072 tokens; Kimi K3 takes 1,048,576.
- Which ranks higher, gpt-oss-20b or Kimi K3?
- Kimi K3 ranks higher on the public arena (gpt-oss-20b #179, Kimi K3 #2). Ranks move as the leaderboard updates.
- Can I use gpt-oss-20b and Kimi K3 with the same API key?
- Yes. On FastMetal's OpenAI-compatible endpoint, switch by passing "gpt-oss-20b" or "kimi-k3" as the model. Each is billed at its own rate from the same prepaid balance.
Try both on one API key
Create an account and add credit to call gpt-oss-20b and Kimi K3 from the browser chat and the API. No monthly fee.