RSI-Exam leaderboard update: GPT-6-astra tops recursive self-improvement benchmark at 0.5126
yuyinzhou_cs · x · 2026-09-19
- New leaderboard update for RSI-Exam, a benchmark testing whether agents can autonomously improve an existing method over long-horizon experiments and generalize to hidden data. Six models added.
- GPT-6-astra (OpenAI) takes #1 at 0.5126, Fable 5.1 (Anthropic) #2 at 0.4813, followed by Qwen3.8 Max (0.3923), Muse Spark 1.3 (0.3882), Gemini 3.8 Flash (0.3406), and Seed Evolving-0909 (0.3243).
- The benchmark spans 88 executable research tasks across 6 domains.
- No model has yet reached the frontier-calibrated reference score of 0.6 — huge headroom remains for recursive self-improvement.
Related event: GPT-6-astra Tops Updated RSI-Exam Leaderboard as Fable 5.1 Debuts Second(3 posts)→
More from Models
- Hindsight ships 0.8B model for agent memory decisions: 2.5x cheaper, 1.9x faster — victor_explore · 2026-11-04
- Mistral Large 4 tech report: 1T-param MoE, NVFP4-quantized to 606GB, runs on 4 GB300 GPUs — _AndrewZhao · 2026-11-04
- Tencent Hunyuan Launches Hy4 with Vision, Claims Top Tier Among Open Models — TencentHunyuan · 2026-11-04
- Claude Code fixes game copyright refusal rule in 2.1.317: naming a game now works — apples_jimmy · 2026-11-04
- Microsoft Phi-4-mini: a 3.8B model with 128K context and 52.8 MMLU-Pro — goyalshaliniuk · 2026-11-04
- Mistral Ministral 3 3B ships in base, instruct and reasoning variants — goyalshaliniuk · 2026-11-04