PhysClosureBench finds frontier agents often miss full physics research closure
Claude-Opus-5 scores 6.29/10 on average but reaches full closure in only 32.7% of attempts.
42
- Mentions · 7d0
- Mentions · 30d3
- All mentions4
- First seenSep 11, 2026
- Last mentioned16d ago
Timeline
4 articles · newest firstMIMESIS: Learning User Simulators as Training Environments for Interactive Agents
MIMESIS, a 9B user simulator, beats cited frontier baselines and improves interactive-agent training.
50
WorldSolver: Can LLM Agents Simulate the Physical Dynamics via Solver Generation?
WorldSolver tests LLM agents on generating physics solvers across 168 tasks from 61 graphics papers.
45
The Self-Expanding Stolen Inference Supply Chain: An AI Agent Harvesting and Re-Serving LLM Access, (Fri, Sep 11th)
An autonomous coding agent harvested LLM API access from poorly secured gateways and aggregated stolen inference capacity behind a self-hosted gateway
68
Appears with
- gpt-5.6-solai model · 2
- glm-5.3ai model · 1
- deepseek-v4-flashai model · 1
- MIMESISai model · 1
- gemini-3.6-flash-highai model · 1
- V2Boardproduct · 1
- SANSvendor · 1
- OpenAIorganization · 1
- New-APIproduct · 1
- PhysClosureBenchproduct · 1
- Sysdigorganization · 1
- GPT-5.5ai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.