Strategy/projects/files/sdr_draco_baseline_analysis.md
+

sdr_draco_baseline_analysis

SDR: DRACO Benchmark Baseline Analysis

Worker auto-analysis, 2026-03-16 11:30 MSK
Pre-evaluation analysis — no API key needed (structural metrics only)


DRACO Score Rankings (LLM-judge, 37 tasks, 0-100 scale)

Rank Model Mean Median Min Max
1 anthropic/claude-opus-4.6 50.3 46.5 8.8 93.2
2 perplexity/sonar-deep-research 35.8 36.2 0.0 70.7
3 google/gemini-3-flash-preview 34.6 33.4 4.9 71.4
4 alibaba/tongyi-deepresearch-30b-a3b 26.7 22.7 0.0 75.2
5 dr-gemini-3-flash 19.7 19.5 0.0 45.7
local_deep_research (SDR) TBD

Gap to beat: ≥50.3 to be competitive with Claude-opus (leader)
Reasonable initial target: ≥35 (beat perplexity/google baselines)


Claude-opus Domain Breakdown (best competitor)

Domain Score N Notes
Academic 69.3/100 12 Strongest — literature review tasks
Medicine 51.5/100 6 Good factual retrieval
General Knowledge 41.4/100 9 Mixed
Technology 34.8/100 10 Weakest — specific tech questions

Implication for SDR: Academic domain is our sweet spot (operator splitting, ML). Technology domain (34.8) = achievable target for SDR first benchmark.


Structural Weaknesses in Claude-opus Baseline

From structural_baseline.json (37 reports):

Metric Value Issue
Diversity score 0.131/1.0 Very low — mainly same-domain sources
Has references section 6/37 (16%) Most reports lack formal references
Word count avg 3,103 Moderate depth
Inline years found 22/37 (59%) Good year diversity in citations
Academic ratio ~0% Surprisingly low for academic papers

Key insight: Even the best competitor (claude-opus) has severe structural weaknesses:
- Only 16% of reports include a references section
- Very low source diversity (0.131/1.0)
- SDR could gain structural score points by ensuring every report has proper references


SDR Improvement Opportunities (pre-API-key analysis)

Quick wins (structural, no quality judgement needed):

  1. Always include references section — Claude-opus fails 84% of the time; SDR could enforce this in prompts
  2. Source diversity — Explicitly request sources from different domains/years in research prompts
  3. Academic ratio — For academic tasks, ensure arxiv/scholar sources are prioritized

LLM quality (needs API key for benchmarking):

  1. Judge model: Anthropic claude-haiku is now configured (eval/config.py); ready for benchmark_2
  2. Academic domain benchmark first: Start with DRACO Academic subset (12 tasks) — fastest path to meaningful comparison

Next Steps for SDR Evaluation

  1. Get ANTHROPIC_API_KEY → provide to Feanor (ask-level task in backlog)
  2. Run benchmark_2 with claude-haiku judge on Academic domain first (12 tasks)
  3. Target score: ≥55 on Academic domain (beat claude-opus 69.3 is aspirational; ≥35 = competitive)
  4. Structural baseline: Already computed for claude-opus; run same for SDR to compare

Eval System Status

  • Tests: 105/105 ✅ (display tests fixed 10:20 today)
  • Judge: Anthropic claude-haiku configured, auto-detect mode
  • Baseline: structural_baseline.json saved for anthropic_claude-opus-4.6
  • Benchmark 2 config: data/benchmarks/benchmark_2.yaml (TBD — needs tasks)
  • Blocker: ANTHROPIC_API_KEY for live benchmark run
Choose icon