opencode-go/longcat-2.5-free
Tested
Longcat 2.5 via OpenCode Go free previews: GSM8K-50 78.0% (39/50, zero failures) through the OpenCode agent harness, so the number is not strictly comparable with the direct-prompted lane. IFEval and SVG are not run for it: the agent harness (large system prompt, unknown temperature) breaks comparability, and SVG probes entered a tool loop with no output in 240 s.
Scores
| Suite | Score |
|---|---|
| GSM8K-50 | 78.0% (39/50) |
Compare with
- opencode-go/longcat-2.5-free vs cohere/north-mini-code:free
- opencode-go/longcat-2.5-free vs dots-studio/dots-3-note-preview:free
- opencode-go/longcat-2.5-free vs kilo-auto/free
- opencode-go/longcat-2.5-free vs liquid/lfm-2.5-2.6b:free
- opencode-go/longcat-2.5-free vs liquid/lfm-3-1.7b:free
- opencode-go/longcat-2.5-free vs nvidia/nemotron-3.5-lightning:free
- opencode-go/longcat-2.5-free vs nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free
- opencode-go/longcat-2.5-free vs nvidia/nemotron-3-ultra-550b-a55b:free
- opencode-go/longcat-2.5-free vs recursive/flint-v1.0:free
Evidence and notes
- Meituan also serves longcat-2.0 paid on OpenRouter; the Go preview is the free route
- Collection paused once on a 429 cooldown, then finished 50/50 after the reset
- SVG probes entered a tool loop (render, verify, refine) with no output in 240 s