gpt-oss-20b is 187.5× cheaper per million tokens (blended). Claude Opus 5 has the larger context window (1,000,000 vs 131,072 tokens, 7.6×). gpt-oss-20b is open-weights; the other is not.
Claude Opus 5 vs gpt-oss-20b
Live catalog fields. Best-in-row is highlighted.
Claude Opus 5
Anthropic · anthropic__claude-opus-5
gpt-oss-20bopen
OpenAI · openai__gpt-oss-20b
Identity
| Field | anthropic__claude-opus-5 | openai__gpt-oss-20b |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Slug | anthropic__claude-opus-5 | openai__gpt-oss-20b |
| Status | live | live |
| Open weights | No | Yes |
| License | — | — |
| HuggingFace | — | openai/gpt-oss-20b |
Quality
| Field | anthropic__claude-opus-5 | openai__gpt-oss-20b |
|---|---|---|
| Quality index (0–100) | 91.5 | — |
Cost
| Field | anthropic__claude-opus-5 | openai__gpt-oss-20b |
|---|---|---|
| Input $ / 1M tokens | $5.00 | $0.03 |
| Cached input $ / 1M | $0.50 | $0.03 |
| Output $ / 1M tokens | $25.00 | $0.13 |
Context
| Field | anthropic__claude-opus-5 | openai__gpt-oss-20b |
|---|---|---|
| Max context tokens | 1,000,000 | 131,072 |
| Max output tokens | 128,000 | 117,964 |
Modalities
| Field | anthropic__claude-opus-5 | openai__gpt-oss-20b |
|---|---|---|
| Input | text, image, file | text |
| Output | text | text |