Anthropic: Claude Opus 4 vs. OpenAI: gpt-oss-20b
Direkter Head-to-Head-Vergleich zweier Frontier-Modelle. OpenAI: gpt-oss-20b gewinnt 3 von 8 Disziplinen.
Letzte Synchronisation:
|
Anthropic: Claude Opus 4
Anthropic
Mehr von Anthropic →
|
OpenAI: gpt-oss-20b
OpenAI
Mehr von OpenAI →
|
|
|---|---|---|
| Quality Index | 25,5 ★ | 14,9 |
| Speed (Tokens/s) | 0 | 187,9 ★ |
| Latency (TTFT) | 0 ms ★ | 569 ms |
| Preis Input (USD/1M) | $15.00 | $0.07 ★ |
| Preis Output (USD/1M) | $75.00 | $0.20 ★ |
| Context Window | — | — |
| Modalitäten | text | text |
| Release | 05/2025 | 08/2025 |
Anthropic: Claude Opus 4
Claude Opus 4 is benchmarked as the world’s best coding model, at time of release, bringing sustained performance on complex, long-running tasks and agent workflows. It sets new benchmarks in...
OpenAI: gpt-oss-20b
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...