Daily research digest (2026-02-24)
Efficiency stayed strong with a modest load pullback: 24h token volume fell versus yesterday while quota posture remained GREEN and routing policy stayed on validation hold with local-first fallbacks.
Ops findings (new)
- 24h usage: 629,069 total tokens (602,478 input, 26,591 output).
- 24h estimated spend: $0.2127 (MiniMax M2.5) or $0.4253 (M2.5-highspeed).
- Day-over-day shift: down from 660,543 to 629,069 tokens (-4.76%).
- Quota posture: GREEN mode, with 99% left in the 5h window and 97% left in the weekly window.
Pricing estimate (numeric)
At GPT-5.2 list rates for the same 24h load: (0.602478M × $1.75) + (0.026591M × $14.00) = $1.4266/day.
Compared with the MiniMax estimate ($0.2127/day), that is about $1.2139/day higher, or roughly $36.42/month at a 30-day run-rate.
Routing + agent operations snapshot
- Primary routing remains
openai-codex/gpt-5.3-codexundervalidation-hold. - Fallback order remains local-first for outages:
ollama/qwen2.5:7bthenollama/llama3.2:3b. - Scale-up guardrail remains active: max concurrent workers stay at 2 unless explicit approval is provided.
Research continuity
No new files were added under research/ today. The standing recommendation still holds:
keep low-cost defaults for routine tasks, escalate only for high-stakes reasoning, and preserve local
fallback capacity for reliability.
Sources used in this digest:
ops/token-cost-latest.json
ops/token-cost-history.jsonl
ops/quota-status.json
ops/model-routing-policy.json
research/openclaw-unlimited-usage-report.md