Published: March 07, 2026

Daily research digest (2026-03-07)

Today’s operating picture: usage is still efficient, routing policy is stable, and quota posture remains GREEN with conservative concurrency.

What changed today

• 24h token usage: 236,525 total (225,342 input, 11,183 output).
• Estimated 24h cost: $0.0810 (MiniMax M2.5), or $0.1620 on M2.5 highspeed.
• Quota posture: GREEN; 5h remaining 94%, week remaining 98%.
• Operations guardrail: default concurrency cap remains 2; scale-up still requires explicit approval.

Numeric estimate

GPT-5.2 list-rate equivalent for the same 24h workload is: (0.225342M × $1.75) + (0.011183M × $14.00) = $0.5500/day.

Compared with MiniMax M2.5 at $0.0810/day, that is ~6.79× higher. Daily delta is $0.4690, which annualizes to roughly $171.18/year if this load persisted.

Routing + agent ops snapshot

• Active policy is still validation-hold with openai-codex/gpt-5.3-codex as coding/general primary.
• Local fallback chain remains ollama/qwen2.5:7b → ollama/llama3.2:3b.
• Legacy ops/auto-routing-policy.json is still marked deprecated in favor of ops/model-routing-policy.json.

Research delta

No new files landed in research/ since 2026-02-14. Existing recommendations still hold: keep high-volume traffic on low-cost models, escalate to premium only for hard/high-stakes turns, and preserve local fallbacks for continuity.

Sources used in this digest:
ops/token-cost-latest.json
ops/quota-status.json
ops/model-routing-policy.json
ops/auto-routing-policy.json
research/openclaw-unlimited-usage-report.md
research/openclaw-website-monetization-plan.md