Agents · Benchmark profile
A long-horizon computer-use benchmark covering realistic workflows across everyday and professional desktop tasks.
Data verified 27 Jul 2026 · Methodology 1.8.0
Benchmark score on OSWorld 2.0
Claude Opus 5 leads at 70.6%, followed by GPT-5.6 Sol (62.6%) and GPT-5.6 Terra (50.2%).
10 ranked models · higher is better · labels show rank and score
| Model | Rank | Provider | Score |
|---|---|---|---|
| Claude Opus 5 | #1 | Anthropic | 70.6% |
| GPT-5.6 Sol | #2 | OpenAI | 62.6% |
| GPT-5.6 Terra | #3 | OpenAI | 50.2% |
| GPT-5.6 Luna | #4 | OpenAI | 45.6% |
| Claude Opus 4.8 | #5 | Anthropic | 20.6% |
| Claude Opus 4.7 (Adaptive) | #6 | Anthropic | 18.2% |
| Muse Spark 1.1 | #7 | Meta | 14.2% |
| Claude Opus 4.7 | #8 | Anthropic | 13.9% |
| GPT-5.5 | #9 | OpenAI | 13% |
| Claude Sonnet 4.6 | #10 | Anthropic | 8.3% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | Claude Opus 5 claude-opus-5 | Anthropic | closed | Reference only | 70.6% |
| #2 | GPT-5.6 Sol gpt-5.6-sol | OpenAI | closed | Reference only | 62.6% |
| #3 | GPT-5.6 Terra gpt-5.6-terra | OpenAI | closed | Estimated reference | 50.2% |
| #4 | GPT-5.6 Luna gpt-5.6-luna | OpenAI | closed | Estimated reference | 45.6% |
| #5 | Claude Opus 4.8 claude-opus-4-8 | Anthropic | closed | Reference only | 20.6% |
| #6 | Claude Opus 4.7 (Adaptive) claude-opus-4-7-max | Anthropic | closed | Estimated reference | 18.2% |
| #7 | Muse Spark 1.1 muse-spark-1-1 | Meta | closed | Reference only | 14.2% |
| #8 | Claude Opus 4.7 claude-opus-4-7 | Anthropic | closed | Reference only | 13.9% |
| #9 | GPT-5.5 gpt-5.5 | OpenAI | closed | Estimated reference | 13% |
| #10 | Claude Sonnet 4.6 claude-sonnet-4-6 | Anthropic | closed | Reference only | 8.3% |
| #11 | Kimi K2.6 kimi-k2.6 | Moonshot AI | open | Estimated reference | 4.6% |
| #12 | MiniMax M3 MiniMax-M3 | MiniMax | open | Reference only | 4.6% |
| #13 | Qwen3.7-Plus qwen3.7-plus | Alibaba Cloud | closed | Reference only | 2.8% |
The top of this snapshot is led by Claude Opus 5 at 70.6%; third place is 20.4 points behind. The top-10 spread is 62.3 points.
About OSWorld 2.0
A long-horizon computer-use benchmark covering realistic workflows across everyday and professional desktop tasks. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A long-horizon computer-use benchmark covering realistic workflows across everyday and professional desktop tasks.
Claude Opus 5 by Anthropic currently leads with 70.6%.
13 models in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related
