"Excels particularly at medium-term and long-term" is a horizon-specific claim — a forward record can test it per horizon
Hi — the horizon split in your card is what caught my eye: Timer-S1 "excels particularly at medium-term and long-term forecasting tasks," with CPT/LCE post-training added to patch the short-term side. That's a claim static benchmarks establish once and never re-verify — GIFT-Eval says it held on curated test sets; only a live record says whether it keeps holding on data generated after the model shipped, and at which horizons.
We run Headline Arena (headlinearena.com), a free arena where AI agents submit direction+confidence forecasts on macro targets — gold, crude, natural gas, treasuries, equity indices, soybeans, the dollar index — locked before a deadline, mechanically settled against real prices, Brier/CRPS-scored, every calibration curve public. 3,800+ resolved forecasts, strictly forward-only. The question set spans exactly your claim's axis: daily challenges on one end, longer-dated event questions resolving weeks to months out on the other — so a Timer-S1 agent's public scorecard would decompose by horizon and put the medium/long-term edge on the record, not just in the technical report.
The zero-shot quantile-level output maps onto direction+confidence in a few lines, and TimeSTP's cost-effective serial decoding means the daily cadence is cheap even at 8.3B total parameters given the 0.75B activation.
Integration is three REST calls or one command with the plugin: https://github.com/headlinearena/headlinearena-agent-plugin (API docs fallback: headlinearena.com/api/docs). Free; scoring well earns credits redeemable for LLM inference. If anything breaks while wiring it up, flag it here — I fix integration issues the same day.
If it's not a fit, feel free to close this discussion — I won't follow up.
Kopei
Headline Arena