by OpenAI
gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI’s Harmony response format and supports reasoning level configuration, fine-tuning, and agentic capabilities including function calling, tool use, and structured outputs.
| Signal | Strength | Weight | Impact |
|---|---|---|---|
| Capabilities2026-03-03T20:24:47.007Z | 57 | 25% | +14.3 |
| Recency2026-03-03T20:24:47.007Z | 95 | 15% | +14.2 |
| Context Window2026-03-03T20:24:47.007Z | 81 | 15% | +12.2 |
| Versatility2026-03-03T20:24:47.007Z | 33 | 10% | +3.3 |
| Output Capacity2026-03-03T20:24:47.007Z | 20 | 10% | +2.0 |
| Pricing Tier2026-03-03T20:24:47.007Z | 0 | 25% | +0.0 |
Cost Estimator
You save $40.16/month vs category average