TinyFish provides free search and fetch, benchmark shows $9.30 per 1,000 tasks
TinyFish, a web‑infrastructure company for AI agents, frames its Search and Fetch APIs as free within published limits – 30 search requests per minute and 500 per hour, plus 150 fetch URLs per minute and 1,000 per day. The company says the free pricing comes from shared infrastructure, specialized web‑native models and the decision not to build a giant index. Model processing still incurs token…
Key points
- TinyFish search free up to 30 requests/min and 500/hour; fetch free up to 150 URLs/min and 1,000/day.
- Benchmark on Sep 28 2026 shows TinyFish total cost $9.30 per 1,000 tasks, lowest among 12 configs, but score 71, the lowest.
- Token usage per page ranges from about 1,184 to 42,774 tokens, consuming up to a third of a 128K-token window.
On the Sep 28, 2026 Artificial Analysis leaderboard, TinyFish Search (web) scored 71 and cost $9.30 per 1,000 benchmark tasks, the lowest total cost among 12 configurations, though it had the lowest score. Model costs across configurations ranged from $8.04 to $21.65 per 1,000 tasks, using GPT‑5.6 Luna (medium) as the fixed answer model. Token usage per page varied widely – from about 1,184 to 42,774 tokens – meaning a single large page can consume up to a third of a 128K‑token window, and three such pages would exceed it.
TinyFish also charges $0.016 per agent step and $0.002 per minute of browser time, with new accounts receiving $8 in wallet funds. The article notes that while retrieval charges are removed, downstream model costs remain, and the benchmark figures should not be confused with raw API pricing.
TinyFish Makes Search Free: What It Really Costs an AI Agent to Read the Web
MarkTechPost · 18 October 2026
Loading the full article…
This text was published by MarkTechPost and written by Asif Razzaq. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Agents & Tools
All →- Six AI agents spam MP office and leak data while chasing £350 train refund · 2 src
- Google DeepMind and Singapore's GovTech develop on-device scam-detecting AI · 1 src
- Anomaly Innovations says breach affected about 1.4% of OpenCode Console users · 2 src
- Claude Code team shares how agents review code with minimal human input · 4 src
- Anthropic is reportedly consulting law firms on negotiation tactics for its AI agents · 2 src
Comments
via GitHub Discussions