Fix horizon bug, multi-provider support, add Sonnet vs Gemini benchmark results

Bug fixes:
- CLI --horizon-years defaulted to 3, silently overriding config presets.
  Now defaults to None so config value (1yr for medium/hard/nightmare) is used.
- Runtime passed a single api_key kwarg regardless of provider, breaking
  Gemini. Now lets LiteLLM resolve keys from provider-specific env vars.
- Removed temperature+top_p from LLM calls (Anthropic rejects both together).
- DB and result filenames now include config name to prevent cross-config collisions.

Benchmark results (1yr horizon, 3 seeds each):
  Sonnet 4.6: medium 2/3, hard 0/3, nightmare 1/3
  Gemini Flash: medium 3/3, hard 1/3, nightmare 1/3
  Gemini has higher win rates (93-98% vs 40-83% on medium).
  Sonnet's ceiling is higher when it survives (nightmare $10.1M vs $478K).

New scripts: plot_comparison.py, plot_sonnet_results.py, notepad_gif.py
Updated README with detailed comparison tables and failure analysis.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

src/yc_bench/cli/__init__.py

@@ -4,6 +4,7 @@ import json
 import sys
 from contextlib import contextmanager
 from decimal import Decimal
+from typing import Optional
 from uuid import UUID
 
 import typer
@@ -80,7 +81,7 @@ app.add_typer(scratchpad_app, name="scratchpad")
 def run_command_cli(
     model: str = typer.Option(..., help="LiteLLM model string (e.g. openrouter/z-ai/glm-5)"),
     seed: int = typer.Option(..., help="Random seed for deterministic world generation"),
-    horizon_years: int = typer.Option(3, help="Simulation horizon in years"),
+    horizon_years: Optional[int] = typer.Option(None, help="Simulation horizon in years (default from config)"),
     company_name: str = typer.Option("BenchCo", help="Name of the simulated company"),
     start_date: str = typer.Option("2025-01-01", help="Simulation start date (YYYY-MM-DD)"),
     config_name: str = typer.Option(