Terminal-Bench
Entity.
- Evidence
- 72 documents · 30 separate publishers
- Standing
- A name from our starting list; its record is kept up to date automatically from coverage
- Time
- first seen 7 Jul 2026, last seen 7 Oct 2026
- Measured
- 19% of mentions say something specific about it; the rest only name it
- As of
- 8 Oct 2026
Where this comes from
- Latent Space · 2026-10-07 · [AINews] Quasi-Riemann-Hypothesis: OpenAI publishes 722 math papers solving 90 of the top 500 open math problems; “the most significant moment” in >100 years of mathematics (opens the source in a new tab)
- Kingy AI · 2026-10-07 · Mistral Large 4 (Le Chonk): Specs, Benchmarks, Pricing and the Open-Weight Promise (opens the source in a new tab)
- SiliconANGLE · 2026-10-07 · Anthropic releases Claude Haiku 5.5 small model and halves Sonnet 5.5 cache read prices (opens the source in a new tab)
- PPC Land · 2026-10-06 · OpenAI text watermark detection drops to 17% after 25% of words change (opens the source in a new tab)
- Tech Times · 2026-10-06 · OpenAI Unveils New AI Text Watermarking System Coming to ChatGPT, Codex, and the API (opens the source in a new tab)
- ArXiv cs.AI · 2026-10-05 · Dissecting model behavior through agent trajectories (opens the source in a new tab)
- Startup Fortune · 2026-10-04 · OpenAI's GPT-6 Astra jumped to 62.7% on AI's hardest reasoning test (opens the source in a new tab)
- Crypto Briefing · 2026-10-04 · Meta, Duke and UC Davis researchers unveil self-improving branches for agent harness optimization (opens the source in a new tab)
- Memeburn · 2026-10-04 · Is Gemini 4 Argon Weaker Than Claude Sonnet 5.5? It Depends Which Scoreboard You Trust (opens the source in a new tab)
- Latent Space · 2026-10-02 · [AINews] Pi 1.0, Pi Durable, and AIE NYC (opens the source in a new tab)
- Business Model Analyst · 2026-10-02 · Google Put an Expiry Date on Argon’s Price. The Model May Expire First. (opens the source in a new tab)
- Latent Space · 2026-10-01 · [AINews] Gemini 4 Argon: GDM’s answer to Astra/Fable, with 1M output (opens the source in a new tab)