@blockrun/clawrouter 0.12.46 → 0.12.47

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/dist/cli.js CHANGED
@@ -1645,36 +1645,36 @@ var DEFAULT_ROUTING_CONFIG = {
1645
1645
  confidenceThreshold: 0.7
1646
1646
  },
1647
1647
  // Auto (balanced) tier configs - current default smart routing
1648
- // Benchmark-tuned 2026-03-16: latency-ranked via blockrun.ai x402 end-to-end
1648
+ // Benchmark-tuned 2026-03-16: balancing quality (retention) + latency
1649
1649
  tiers: {
1650
1650
  SIMPLE: {
1651
- primary: "xai/grok-4-fast-non-reasoning",
1652
- // 1,143ms, $0.20/$0.50fastest overall
1651
+ primary: "google/gemini-2.5-flash",
1652
+ // 1,238ms, 60% retention (best) fast AND quality
1653
1653
  fallback: [
1654
- "xai/grok-3-mini",
1655
- // 1,202ms, $0.30/$0.50
1656
- "google/gemini-2.5-flash",
1657
- // 1,238ms, 60% retention (best)
1658
- "google/gemini-2.5-flash-lite",
1659
- // 1,353ms, 1M context, ultra cheap ($0.10/$0.40)
1660
1654
  "deepseek/deepseek-chat",
1661
1655
  // 1,431ms, 41% retention
1656
+ "moonshot/kimi-k2.5",
1657
+ // 1,646ms, strong quality
1658
+ "google/gemini-2.5-flash-lite",
1659
+ // 1,353ms, 1M context, ultra cheap ($0.10/$0.40)
1660
+ "xai/grok-4-fast-non-reasoning",
1661
+ // 1,143ms, $0.20/$0.50 — fast fallback
1662
1662
  "nvidia/gpt-oss-120b"
1663
1663
  // 1,252ms, FREE fallback
1664
1664
  ]
1665
1665
  },
1666
1666
  MEDIUM: {
1667
- primary: "xai/grok-4-1-fast-non-reasoning",
1668
- // 1,244ms, $0.20/$0.50fast + tool calling
1667
+ primary: "moonshot/kimi-k2.5",
1668
+ // 1,646ms, $0.60/$3.00strong tool use, quality output
1669
1669
  fallback: [
1670
1670
  "deepseek/deepseek-chat",
1671
1671
  // 1,431ms, 41% retention
1672
- "moonshot/kimi-k2.5",
1673
- // 1,646ms, strong tool use quality
1674
1672
  "google/gemini-2.5-flash",
1675
1673
  // 1,238ms, 60% retention
1676
1674
  "google/gemini-2.5-flash-lite",
1677
1675
  // 1,353ms, 1M context ($0.10/$0.40)
1676
+ "xai/grok-4-1-fast-non-reasoning",
1677
+ // 1,244ms, fast fallback
1678
1678
  "xai/grok-3-mini"
1679
1679
  // 1,202ms, $0.30/$0.50
1680
1680
  ]
@@ -1805,15 +1805,15 @@ var DEFAULT_ROUTING_CONFIG = {
1805
1805
  // Agentic tier configs - models that excel at multi-step autonomous tasks
1806
1806
  agenticTiers: {
1807
1807
  SIMPLE: {
1808
- primary: "xai/grok-4-1-fast-non-reasoning",
1809
- // 1,244ms, $0.20/$0.50 fast tool calling
1808
+ primary: "openai/gpt-4o-mini",
1809
+ // $0.15/$0.60 - best tool compliance at lowest cost
1810
1810
  fallback: [
1811
- "openai/gpt-4o-mini",
1812
- // 2,764ms, $0.15/$0.60 - reliable tool compliance
1813
1811
  "moonshot/kimi-k2.5",
1814
1812
  // 1,646ms, strong tool use quality
1815
- "anthropic/claude-haiku-4.5"
1813
+ "anthropic/claude-haiku-4.5",
1816
1814
  // 2,305ms
1815
+ "xai/grok-4-1-fast-non-reasoning"
1816
+ // 1,244ms, fast fallback
1817
1817
  ]
1818
1818
  },
1819
1819
  MEDIUM: {