Written by Jakub Rusinowski · Last updated June 15, 2026
OpenAI's open-weight 120B dense model released April 2026 under Apache 2.0 — their first substantial open-source release since GPT-2. Matches GPT-4o on most benchmarks at 65 GB Q4, making it the highest-quality open model at the workstation tier. Signals OpenAI's strategic shift toward supporting on-premise enterprise deployments.
| GPT-oss 120B | Min 65 GB VRAM · Q4_K_M · 128,000 ctx · ollama run gpt-oss:120b |
Install Ollama then run: ollama run gpt-oss:120b
Minimum VRAM: 65 GB. For best results use Q4_K_M quantization.
GPT-oss 120B needs about 65 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: GPT-oss 120B (65 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
GPT-oss 120B's smallest variant needs about 65 GB, which exceeds a single RTX 4090 (24 GB). Use multiple GPUs, a higher-VRAM card, or Apple Silicon with large unified memory.
Q4_K_M is the best balance of quality and VRAM for GPT-oss 120B in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
Install Ollama, then run: ollama run gpt-oss:120b. This downloads GPT-oss 120B and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.