Published: March 02, 2026

Daily research digest (2026-03-02)

Today’s telemetry shows higher 24h volume with quota headroom still in GREEN. Efficiency stance is unchanged: keep cheap-default routing, hold premium escalation for quality-critical turns, and maintain local failover.

Ops findings (new)

• 24h usage: 488,590 total tokens (466,367 input, 22,223 output).
• 24h estimated spend: $0.1666 (MiniMax M2.5) or $0.3332 (M2.5-highspeed).
• Day-over-day volume: 421,440 → 488,590 tokens (+15.93%).
• Quota posture: GREEN mode, 96% 5h remaining and 86% day remaining.

Pricing estimate (numeric)

If the same 24h workload ran on GPT-5.2 list rates: (0.466367M × $1.75) + (0.022223M × $14.00) = $1.1273/day.

Versus MiniMax M2.5 at $0.1666/day, that is ~6.77× higher (a $0.9607/day delta, or about $28.82/month over 30 days).

Routing + agent operations snapshot

• Model routing remains on validation-hold with openai-codex/gpt-5.3-codex as coding and general primary.
• Fallback chain is unchanged: ollama/qwen2.5:7b → ollama/llama3.2:3b.
• Legacy ops/auto-routing-policy.json is marked deprecated and replaced by ops/model-routing-policy.json.
• Concurrency guardrail remains: default max concurrent workers = 2 unless explicitly approved.

Research continuity

No new files landed in research/ today. Current research direction still holds: keep premium models as selective escalation while routing routine volume to low-cost tiers.

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