router.json

{
 "version": "v1.2-traffic",
 "kind": "trained",
 "hf_repo": "experiential-labs/coding-router",
 "embed_model_mlx": "encoder-mlx-4bit",
 "embed_model_torch": "encoder-fp16",
 "n_tasks": 207,
 "n_repos": 88,
 "embed_backend": "local",
 "T": 0.05453708011655199,
 "lam": 0.1,
 "sim_floor": 0.4699,
 "fallback_arm_index": 5,
 "arms": [
  "fable5_xhigh",
  "luna_high",
  "luna_low",
  "luna_max",
  "luna_medium",
  "opus5_high",
  "sol_xhigh",
  "sonnet5_high",
  "terra_high",
  "terra_max"
 ],
 "arm_spec": {
  "fable5_xhigh": {
   "model": "claude-fable-5",
   "effort": "xhigh",
   "provider": "anthropic",
   "request_kwargs": {
    "model": "claude-fable-5",
    "thinking": {
     "type": "adaptive"
    },
    "output_config": {
     "effort": "xhigh"
    }
   }
  },
  "luna_high": {
   "model": "gpt-5.6-luna",
   "effort": "high",
   "provider": "openai",
   "request_kwargs": {
    "model": "gpt-5.6-luna",
    "reasoning": {
     "effort": "high"
    }
   }
  },
  "luna_low": {
   "model": "gpt-5.6-luna",
   "effort": "low",
   "provider": "openai",
   "request_kwargs": {
    "model": "gpt-5.6-luna",
    "reasoning": {
     "effort": "low"
    }
   }
  },
  "luna_max": {
   "model": "gpt-5.6-luna",
   "effort": "max",
   "provider": "openai",
   "request_kwargs": {
    "model": "gpt-5.6-luna",
    "reasoning": {
     "effort": "max"
    }
   }
  },
  "luna_medium": {
   "model": "gpt-5.6-luna",
   "effort": "medium",
   "provider": "openai",
   "request_kwargs": {
    "model": "gpt-5.6-luna",
    "reasoning": {
     "effort": "medium"
    }
   }
  },
  "opus5_high": {
   "model": "claude-opus-5",
   "effort": "high",
   "provider": "anthropic",
   "request_kwargs": {
    "model": "claude-opus-5",
    "thinking": {
     "type": "adaptive"
    },
    "output_config": {
     "effort": "high"
    }
   }
  },
  "sol_xhigh": {
   "model": "gpt-5.6-sol",
   "effort": "xhigh",
   "provider": "openai",
   "request_kwargs": {
    "model": "gpt-5.6-sol",
    "reasoning": {
     "effort": "xhigh"
    }
   }
  },
  "sonnet5_high": {
   "model": "claude-sonnet-5",
   "effort": "high",
   "provider": "anthropic",
   "request_kwargs": {
    "model": "claude-sonnet-5",
    "thinking": {
     "type": "adaptive"
    },
    "output_config": {
     "effort": "high"
    }
   }
  },
  "terra_high": {
   "model": "gpt-5.6-terra",
   "effort": "high",
   "provider": "openai",
   "request_kwargs": {
    "model": "gpt-5.6-terra",
    "reasoning": {
     "effort": "high"
    }
   }
  },
  "terra_max": {
   "model": "gpt-5.6-terra",
   "effort": "max",
   "provider": "openai",
   "request_kwargs": {
    "model": "gpt-5.6-terra",
    "reasoning": {
     "effort": "max"
    }
   }
  }
 },
 "provenance": "EXP-012 winning recipe `reward_lcb_b0.2`: Qwen3-Embedding-0.6B with LoRA r=16 (q/k/v/o projections) trained by exact expected reward -E_pi[graded - 3.0*cost] on the LiveCodeBench 7x76 matrix, KL(pi||pi_init) anchor beta=0.2, 150 steps, checkpoint step 50 (T=0.0545), lam=0.01 -- (step, lam) selected ONLY on the six standard seeds' inner 75/25 DeepSWE-train splits (feasibility inner_graded >= inner-best-arm - 0.01, then max mean cost ratio; feasible on 5/6 inner splits). Memory/calibration side: the full 110-task DeepSWE v1.1 evidence (41 arms x 110 tasks, 88 repos), bank embedded with the merged fp16 encoder via sentence-transformers. Sweep evidence (per-seed selection, 6 seeds): selected-holdout mean graded 0.9484 at $60.17/split -- +0.015 graded and 2.1x cheaper than the deployable always-best-train baseline (0.9336/$126.28); parity, NOT better, vs the hindsight-best static arm. Cost figures are matrix-based (June 2026 collection) and pending live re-benchmark per EXP-014 (live drift measured). | v1.1 REBUILD (2026-08-01): bank cells replaced with the EXP-015 LIVE DeepSWE matrix (1,130 trials, dense 10 arms x 113 tasks; the 3 tasks without bank texts dropped -> 10x110) because the published June matrix is quality-stale live (e.g. luna_max 0.946 published vs 0.687 live). Roster 41 -> 10 live arms; med_cost/fallback from live cells (opus5_high, live f2p 0.951); tuned encoder, bank embeddings and T unchanged; lambda re-selected on inner train-side splits only (0.005). | v1.2-traffic (2026-08-01): union bank = DeepSWE-live 110 (unchanged cells) + 65 LCB-live + 32 utility tasks, all arms measured LIVE this lane (LCB via the E2B harness on the 92-task covered subset; utility = 32 short interactive asks with deterministic verifiers on 4 pinned OSS repos). graded = harness fraction; med_cost per arm = median of live per-task costs pooled across the three sources. sim_floor = LOO-nearest p10 over the union bank. lam re-selected train-side only under a 50/30/20 utility/LCB/DeepSWE traffic weighting.",
 "scope_warning": "Validated on repo-issue statements (DeepSWE), competitive-programming agent tasks (LCB) and SHORT interactive asks (utility suite: bugfix/feature/refactor/question, 40-250 chars) \u2014 short prompts now land in-distribution instead of abstaining. Abstention (escalate to strongest arm) is reserved for text unlike any bank task; always check Decision.off_distribution.",
 "selection": {
  "lam_grid": [
   {
    "lam": 0.001,
    "feas": 6,
    "mean_ratio": 1.081393772849211,
    "mean_inner_graded": 0.7660379605122892
   },
   {
    "lam": 0.002,
    "feas": 6,
    "mean_ratio": 1.1000667545641503,
    "mean_inner_graded": 0.7659905447036618
   },
   {
    "lam": 0.005,
    "feas": 6,
    "mean_ratio": 1.2811509404979988,
    "mean_inner_graded": 0.765745203615449
   },
   {
    "lam": 0.01,
    "feas": 6,
    "mean_ratio": 1.7326780876087458,
    "mean_inner_graded": 0.7656267731370289
   },
   {
    "lam": 0.02,
    "feas": 6,
    "mean_ratio": 2.452558818623846,
    "mean_inner_graded": 0.7645313997014641
   },
   {
    "lam": 0.05,
    "feas": 6,
    "mean_ratio": 3.2512255254226896,
    "mean_inner_graded": 0.7644669476479534
   },
   {
    "lam": 0.1,
    "feas": 6,
    "mean_ratio": 3.5954684774287116,
    "mean_inner_graded": 0.7617608602098715
   },
   {
    "lam": 0.2,
    "feas": 5,
    "mean_ratio": 3.9614547647054508,
    "mean_inner_graded": 0.7608279933293333
   },
   {
    "lam": 0.5,
    "feas": 5,
    "mean_ratio": 4.002295380963825,
    "mean_inner_graded": 0.7593800766626666
   },
   {
    "lam": 1.0,
    "feas": 5,
    "mean_ratio": 4.061556689358738,
    "mean_inner_graded": 0.7589185461017406
   }
  ],
  "selected_lam": 0.1,
  "policy": "6 seeded group-stratified splits, inner train-side only; metric weights sources 50% utility / 30% LCB / 20% DeepSWE (assumed traffic mix, documented); encoder/T frozen"
 }
}