No model among 23 reaches 80% threshold on Vals AI and NextLadder Access to Justice Benchmark
NextLadder Ventures and Vals AI announced the Access to Justice Benchmark on November 10, 2026. None of the 23 models tested reached the 80% threshold for a reliable first point of contact across 252 legal-aid scenarios.
NextLadder Ventures and Vals AI announced the Access to Justice Benchmark on November 10, 2026. None of the 23 AI models tested reached the benchmark's 80% threshold for a reliable first point of contact across 252 legal-aid scenarios, according to the announcement.
Results
The results show a sharp divide between explaining law and finding practical help, as reported by Vals AI:
- General legal explanations: models satisfied 78% to 87% of expert criteria.
- Identifying specific resources: models satisfied only 30% to 50% of expert criteria.
Benchmark design
- Models tested: 23
- Scenarios: 252 legal-aid scenarios
- Pass threshold: 80%, set as the bar for a reliable first point of contact
- Organizations: NextLadder Ventures and Vals AI
Per-model scores, pricing and context window details were not available in the summary of the announcement.
What this means
The 80% threshold is the benchmark's own bar for a reliable first point of contact, and none of the 23 models met it. According to Vals AI, the gap is not in explaining the law: the reported 78% to 87% range for general explanations sits much closer to that bar than the 30% to 50% range for identifying specific resources. For teams building legal-aid tools on these models, the reported results point to resource identification as the weaker area to test and address.
Related Articles
Vals AI finds 316 expletives in Thinking Machines' Inkling 2 Small reasoning logs across 2,655 benchmark runs
Vals AI analyzed the chain-of-thought logs of Thinking Machines' Inkling 2 Small across 2,655 runs on 13 benchmarks and found expletives in 50 runs, totaling 316 instances. Two runaway runs produced 241 of them, and swearing did not work as a reliable failure signal.
Anthropic's Fable 5.2 tops Remote Labor Index at 41.3%, completing real freelance jobs at professional quality
Anthropic's Fable 5.2 scores 41.3 percent on the Remote Labor Index from Scale Labs and the Center for AI Safety, which measures how often AI agents complete real freelance jobs well enough for a client to accept. The next-best systems, Opus 5.5 and GPT 6 Astra, score 21.25 and 20.83 percent. The top score at the benchmark's launch about a year ago was 2.5 percent.
Co-Design Bench: best coding agent fully solves only 26.4% of real user change requests
Researchers from Alibaba and several Asian universities released Co-Design Bench, a coding-agent test built from real user change requests. According to the paper, the best of ten models, GPT-6 Astra, meets about 70 percent of a task's requirements on average but is fully correct on only 26.4 percent of tasks.
Vals AI: OpenAI's GPT-6 Sol Daybreak hits 80.8% on CyberOffense, open-weight Deepseek V4.1 Flash reaches 72.8%
On Vals AI's CyberOffense benchmark, OpenAI's restricted GPT-6 Sol Daybreak completed 80.8 percent of objectives. The open-weight Deepseek V4.1 Flash 1017 reached 72.8 percent at about $3 per run instead of roughly $22.
Comments
Loading...