Skip to content

Claude Haiku 5.5 (effort default (no --effort flag) · thinking on) on Build a full game (code, art, audio) with AI

AClaude Haiku 5.5default · thinking on · Claude Code 2.1.299 (headless -p) · cloud · 2026-10-16 · by @bobsiopener

Trust 80%Fully autonomoussuccess · first try
Estimated cost
$0.05
Wall time
10m
Human step-ins
0
Outcome
Pass

Compare with other models on this task →

Summary

Exact prompt

Game jam, scaled down: in one pass build a complete small arcade game as a single file index.html (no external assets or libraries). Theme: "a capybara crossing a river". Must include: a title screen with Start, keyboard controls, procedurally drawn art on a <canvas> (no image files), at least 2 sound effects and background music generated with the Web Audio API (start audio on user gesture), a score, increasing difficulty, a game-over screen with restart, and a high score saved in localStorage.

Task prompt given to the agent (headless, offline, synthetic fixture inputs): Game jam, scaled down: in one pass build a complete small arcade game as a single file index.html (no external assets or libraries). Theme: "a capybara crossing a river". Must include: a title screen with Start, keyboard controls, procedurally drawn art on a <canvas> (no image files), at least 2 sound effects and background music generated with the Web Audio API (start audio on user gesture), a score, increasing difficulty, a game-over screen with restart, and a high score saved in localStorage. Inspired by https://x.com/qusaxz/status/2077427986689843623 (unverified; not checked with the author). This is an independent re-run on a reproducible synthetic task; it does not report or reproduce the original post's numbers. Cost shown is a list-price API-equivalent estimate (run on a Claude Pro subscription, not billed).

Success criteria: All automated rubric checks pass: single file, no external libs/assets; canvas drawing; Web Audio; oscillator music/sfx; title/start screen; game over + restart; localStorage high score; capybara theme; no uncaught console errors.

Setup notes: Cost is an ESTIMATE: Claude Code's list-price API-equivalent figure. The run used a Claude Pro subscription, so nothing was billed. input_tokens = all input incl. cache-creation and cache-read tokens; cached_tokens = cache reads (subset of input). Claude Code's cost bills cache writes at the 1h cache-write premium, so it can differ from a flat input-price recompute. Web disabled; tools Read/Write/Edit/Glob/Grep/Bash; single run; scored automatically (9/9 checks).

Estimated cost List-price estimate; not actually billed.

Metrics

Estimated cost
$0.05
Wall time
10m
Human step-ins
0
Turns
3
Model usage
ModelInput tokOutput tokCost
AClaude Haiku 5.5anthropic/claude-haiku-5.5effort default (no --effort flag) · thinking on126,90571,211$0.05

Reproduce this run

  1. Use the use case prompt on the use case page.
  2. Same setup: Claude Haiku 5.5 (anthropic/claude-haiku-5.5, effort default (no --effort flag) · thinking on) via Claude Code 2.1.299 (headless -p), cloud.
  3. It counts as a success only if it meets every success criterion: All automated rubric checks pass: single file, no external libs/assets; canvas drawing; Web Audio; oscillator music/sfx; title/start screen; game over + restart; localStorage high score; capybara theme; no uncaught console errors.
  4. Log model, tokens, wall time, human step-ins and the evidence log, then submit it as a reproduction. A different model? Submit a head-to-head rerun.

How runs are scored: Methodology.

Was this useful?

⇄ Rerun on another model

Automated checks

Trust 80% comes from these checks (how Trust % works).

Evidence log
{
 "duration_api_ms": 518408,
 "stop_reason": "end_turn",
 "session_id": "[REDACTED]",
 "total_cost_usd": 0.05147127,
 "usage": {
  "input_tokens": 6,
  "cache_creation_input_tokens": 76822,
  "cache_read_input_tokens": 50077,
  "output_tokens": 71211,
  "output_tokens_details": {
   "thinking_tokens": 58262
  },
  "server_tool_use": {
   "web_search_requests": 0,
   "web_fetch_requests": 0
  },
  "service_tier": "standard",
  "cache_creation": {
   "ephemeral_1h_input_tokens": 76822,
   "ephemeral_5m_input_tokens": 0
  },
  "inference_geo": "global",
  "iterations": [
   {
    "input_tokens": 2,
    "output_tokens": 755,
    "cache_read_input_tokens": 18769,
    "cache_creation_input_tokens": 70592,
    "cache_creation": {
     "ephemeral_5m_input_tokens": 0,
     "ephemeral_1h_input_tokens": 70592
    },
    "type": "message"
   }
  ],
  "speed": "standard",
  "fallback_credit": null
 },
 "modelUsage": {
  "claude-haiku-5-5": {
   "inputTokens": 6,
   "outputTokens": 71211,
   "cacheReadInputTokens": 50077,
   "cacheCreationInputTokens": 76822,
   "webSearchRequests": 0,
   "costUSD": 0.05147127,
   "contextWindow": 1000000,
   "maxOutputTokens": 128000,
   "thinkingTokens": 58262,
   "canonicalModel": "claude-haiku-5-5",
   "provider": "firstParty",
   "costBasis": "list"
  }
 },
 "permission_denials": [],
 "terminal_reason": "completed",
 "fast_mode_state": "off",
 "fast_mode_disabled_reason": "sdk_opt_in_required",
 "subagent_stats": {
  "spawned": 0,
  "requested": {
   "background": 0,
   "foreground": 0,
   "unset": 0
  },
  "started_in_background": 0,
  "max_depth": 0,
  "spawned_by_subagents": 0,
  "completed": 0,
  "failed": 0,
  "killed": {
   "parent": 0,
   "user": 0,
   "system": 0
  },
  "refused": {
   "depth_limit": 0,
   "concurrency_limit": 0,
   "budget": 0
  },
  "by_type": {}
 },
 "safety_stops": 0,
 "is_error": false,
 "num_turns": 3,
 "subtype": "success",
 "api_error_status": null,
 "result": "I built the game as a single file: `./usecase-runner/runs/17-haiku/work/index.html`. The script passes a Node syntax check, but I have not run it in a browser, so gameplay balance and rendering are unverified.\n\n**How it plays**\n- You hop a capybara up a river, one tile per press, using the arrow keys or WASD.\n- Rows alternate between water with drifting logs and grass banks. Ride a log, or drown if it slides out from under you. Logs carry you off-screen if you ride them past the edge.\n- The goal bank has six dens. Each filled den is worth 100 points, and each new row reached is worth 10. Filling all six clears the level for 500 points.\n- You have 3 lives. Clearing a level raises the lane speed by 15% and shortens the logs, up to a cap of 2.2\u00d7.\n- Game over shows the score, high score and a restart option.\n\n**Requirements**\n- **Title screen:** a Start button, with Enter, Space or a click to start.\n- **Controls:** arrows/WASD to hop, P or Esc to pause, M to toggle music.\n- **Art:** everything is drawn procedurally on the canvas (capybara, logs, water, bushes, dens, splash ripples and particles). No image files or libraries.\n- **Audio:** the AudioContext is created on the first key press or click. Sound effects are hop, splash, den and level-up, plus a game-over tone. Background music is a looping lead, bass and kick generated with oscillators.\n- **Score and difficulty:** score and level are shown in the HUD. Difficulty increases per level as described above.\n- **High score:** saved to `localStorage` under `capybara-crossing-highscore`. Storage errors are ignored, so the game still runs if storage is blocked.\n\n**Choices I made that you may want to change**\n- Each game has 3 lives rather than instant game over on the first splash.\n- Pressing a direction during a hop buffers one input, so quick presses are not lost.\n- The game auto-pauses when the tab is hidden.\n- The title screen shows the lane scene with the logs drifting.\n\nThis is not a git repository, so I did not commit anything.",
 "ttft_ms": 479945,
 "type": "result",
 "duration_ms": 518792,
 "uuid": "[REDACTED]",
 "ttft_stream_ms": 1053,
 "time_to_request_ms": 213,
 "first_content_frame_ms": 1053,
 "queued_turn_count": 0,
 "result_index": 0
}

Get new runs and ranking changes weekly