OpenAI
OpenAI: gpt-oss-120b
openai/gpt-oss-120b
gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases.
- Context window: 131,072 tokens
- Input: text
- Output: text
- Pricing: $0.037/M input tokens, $0.17/M output tokens
Community sentiment
Positive (+86) — 8 positive, 4 negative, 3 neutral across 15 community samples from X and the web, as of 2026-10-10.
Conversation is mostly positive about gpt-oss-120b’s accessibility and performance — users praise free/local access, ability to run on consumer hardware, and solid benchmark/cost-efficiency. Critics argue the model is dated compared with newer smaller models and complain about pricing differences across vendors.
- “I JUST RAN GPT OSS 120B FOR $0 turned on Antseed no account needed OpenCode pointed to localhost:8377 sent “hello” GPT OSS 120B answered cost me $0 how it works: • local OpenAI-compatible endpoint • works with OpenCode, Claude Code, Cursor, Hermes and more • pick a seller and mod” — X
- “ran this through mbench. it scored below gpt-oss-120b. https://t.co/Zkb4828TNr” — X
- “@fabrice_mayrand Its already more expensive than openrouter On your product the gpt-oss-120b is priced 0.04/0.2 but on openrouter its 0.03/0.17 per m token in/out” — X
- “sorry but gpt-oss-120B is dated. no one calls it a "major [..] for reasoning and agentic workloads". good job for EU on producing a model but lets not muddy the waters. https://t.co/2O49Cby0N2” — X
View on OpenRouter.
Model data sourced from OpenRouter.