OpenAI's gpt-oss Models Match Premium Power, No Cloud Needed
Main Idea
OpenAI has launched two new AI language models, gpt-oss-120b and gpt-oss-20b, which support long inputs up to 128,000 tokens and perform similarly to the GPT-4o model.
Key Points
1. OpenAI introduced two new models: gpt-oss-120b (requires 80GB graphics card) and gpt-oss-20b (needs 16GB memory).
2. Both models support inputs up to 128,000 tokens, matching GPT-4o's limit.
3. The models are available under the Apache 2.0 license, allowing personal and commercial use.
4. They perform comparably to OpenAI’s o4-mini model in reasoning tasks.
5. Developers can adjust processing power usage via system messages.
Description
OpenAI has launched two new artificial intelligence (AI) language models that are free to use and can run directly on consumer computers.
Latest News
- Coinbase Hit by $300,000 Loss From Token Approval Mistake2025-08-15 08:58:06
- $200 Million Paynet Coin Ponzi Busted in Vietnam, $38 Million in Assets Seized2025-08-15 08:55:38
- Justin Sun Fights Bloomberg’s Plan to Expose His Crypto Fortune2025-08-14 10:09:44
- a16z and DeFi Education Fund Push SEC for Safe Harbor on NFTs and DeFi2025-08-14 04:51:39
- Google Play Tightens Crypto Wallet Rules, Spares Self-Custody Apps2025-08-14 04:51:07