gpt-oss-120b is 3.6× cheaper per million tokens (blended). gpt-oss-120b has the larger context window (131,072 vs 80,000 tokens, 1.6×).
Llama 3.2 3B Instruct vs gpt-oss-120b
Live catalog fields. Best-in-row is highlighted.
Llama 3.2 3B Instructopen
Meta · meta-llama__llama-3.2-3b-instruct
gpt-oss-120bopen
OpenAI · openai__gpt-oss-120b
Identity
| Field | meta-llama__llama-3.2-3b-instruct | openai__gpt-oss-120b |
|---|---|---|
| Provider | Meta | OpenAI |
| Slug | meta-llama__llama-3.2-3b-instruct | openai__gpt-oss-120b |
| Status | live | live |
| Open weights | Yes | Yes |
| License | — | — |
| HuggingFace | meta-llama/Llama-3.2-3B-Instruct | openai/gpt-oss-120b |
Quality
| Field | meta-llama__llama-3.2-3b-instruct | openai__gpt-oss-120b |
|---|---|---|
| Quality index (0–100) | — | — |
Cost
| Field | meta-llama__llama-3.2-3b-instruct | openai__gpt-oss-120b |
|---|---|---|
| Input $ / 1M tokens | $0.05 | $0.02 |
| Cached input $ / 1M | — | $0.02 |
| Output $ / 1M tokens | $0.34 | $0.09 |
Context
| Field | meta-llama__llama-3.2-3b-instruct | openai__gpt-oss-120b |
|---|---|---|
| Max context tokens | 80,000 | 131,072 |
| Max output tokens | 72,000 | 65,536 |
Modalities
| Field | meta-llama__llama-3.2-3b-instruct | openai__gpt-oss-120b |
|---|---|---|
| Input | text | text |
| Output | text | text |