CUA-bench leaderboard
As of 6 Nov 2026, the top score on CUA-bench is 19.2% by GPT-6 Astra (OpenAI), ahead of Claude Opus 5.5 at 14.0%, across 9 model configurations with a published result (Vals AI, result dated 13 Oct 2026). A computer-use agent is put in front of six commercial games with screen, keyboard and mouse and scored on progress. Three games are public, three held out. CUA-bench is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- GPT-6 Astra holds the top CUA-bench score at 19.2% (Vals AI, 13 Oct 2026).
- Claude Opus 5.5 is second on CUA-bench at 14.0% (Vals AI, 13 Oct 2026).
- CUA-bench: Playing six commercial video games with only a keyboard and a mouse; half are held out. Run by Vals AI (Vals AI).
- 9 model configurations have a published CUA-bench result on BenchLeader (BenchLeader, 6 Nov 2026).
- CUA-bench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Playing six commercial video games with only a keyboard and a mouse; half are held out. Run by Vals AI.
- Published by
- Vals AI
- Category
- Agents & tools
- Index weight
- Reference only
- Models
- 9
- Data as of
- 6 Nov 2026
Vals AI (vals.ai).
What the test looks like
A computer-use agent is put in front of six commercial games with screen, keyboard and mouse and scored on progress. Three games are public, three held out.
How it is scored
Progress score across games, run by Vals AI.
What to keep in mind
Games test perception and control loops more than knowledge; scaffolding and frame rate matter.
- 1GPT-6 Astra (max)19.2%
- 2Claude Opus 5.5 (max)14.0%
- 3Claude Fable 5.1 (max)13.2%
- 4Claude Opus 5 (max)9.0%
- 5GPT-5.6 Sol (max)8.3%
- 6Muse Wonder (max)7.2%
- 7Muse Spark 1.3 (max)5.8%
- 8Gemini 4 Argon (high)4.8%
- 9Gemini 3.8 Flash (high)4.2%
| # | |||||
|---|---|---|---|---|---|
| 1 |  GPT-6 Astramax | 19.2% | 67.8 | $1821.37 | 2026-10-13 |
| 2 |  Claude Opus 5.5max | 14.0% | 69.3 | $188.04 | 2026-10-13 |
| 3 | Claude Fable 5.1max | 13.2% | 67.2 | $323.78 | 2026-10-13 |
| 4 | Claude Opus 5max | 9.0% | 65.4 | $232.23 | 2026-10-13 |
| 5 | GPT-5.6 Solmax | 8.3% | 65.4 | $206.99 | 2026-10-13 |
| 6 |  Muse Wondermax | 7.2% | 62.3 | $597.21 | 2026-10-13 |
| 7 | Muse Spark 1.3max | 5.8% | 62.0 | $143.48 | 2026-10-13 |
| 8 |  Gemini 4 Argonhigh | 4.8% | 66.0 | $193.78 | 2026-10-13 |
| 9 | Gemini 3.8 Flashhigh | 4.2% | 62.0 | $280.03 | 2026-10-13 |
More agents & tools benchmarks · All LLM benchmarks
Related benchmarks
Top head-to-heads on CUA-bench
Cite as: BenchLeader, “CUA-bench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vals_cua_bench.
CUA-bench: questions
- What does CUA-bench measure?
- A computer-use agent is put in front of six commercial games with screen, keyboard and mouse and scored on progress. Three games are public, three held out. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on CUA-bench?
- As of 6 Nov 2026, the highest score on CUA-bench is 19.2% by GPT-6 Astra (result dated 13 Oct 2026), ahead of Claude Opus 5.5 at 14.0%. Scores come from Vals AI and are re-read daily.
- Which AI model leads CUA-bench?
- GPT-6 Astra leads CUA-bench with 19.2% as of 6 Nov 2026 (result dated 13 Oct 2026), ahead of Claude Opus 5.5 at 14.0%.
- How many models have CUA-bench results?
- 9 model configurations have a CUA-bench result on BenchLeader, all taken from Vals AI.
- Who runs CUA-bench and how often is it updated?
- CUA-bench is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does CUA-bench count toward the BenchLeader Index?
- No. CUA-bench is shown for reference but left out of the composite index.