can1357/oh-my-pi · error

--prompt requires --profile chat or generation

Error message

--prompt requires --profile chat or generation

What it means

A custom --prompt only makes sense for workloads that use a single user prompt (chat or generation profiles). The mix profile builds its own multi-segment workload and prefill uses prefix prefill bytes, so --prompt would be ignored; the CLI rejects the combination (only when not in --cache mode, which has its own prompt rules).

Source

Thrown at packages/coding-agent/src/cli/bench-cli.ts:879

		throw new Error("Use --cache-pairs instead of --runs with --cache");
	if (cacheMode && command.flags.prompt !== undefined) throw new Error("--cache builds its own stable-prefix prompts");
	if (cacheMode && command.flags.profile !== undefined) throw new Error("--profile cannot be combined with --cache");
	if (cacheMode && (command.flags.par ?? 1) > 1) {
		throw new Error("--par cannot parallelize cold/warm pairs; use --cache-concurrency instead");
	}
	const profileFlag = command.flags.profile;
	if (
		profileFlag !== undefined &&
		profileFlag !== "mix" &&
		profileFlag !== "chat" &&
		profileFlag !== "prefill" &&
		profileFlag !== "generation"
	) {
		throw new Error(`Unknown --profile "${profileFlag}" (expected mix, chat, prefill, or generation)`);
	}
	const profile: BenchProfile = profileFlag ?? "mix";
	if (!cacheMode && command.flags.prompt !== undefined && profile !== "chat" && profile !== "generation") {
		throw new Error("--prompt requires --profile chat or generation");
	}
	if (command.flags.prefillBytes !== undefined && (cacheMode || (profile !== "mix" && profile !== "prefill"))) {
		throw new Error("--prefill-bytes requires prefill challenges (--profile mix or prefill)");
	}

	const cachePairs = cacheMode
		? normalizePositiveInteger("cache-pairs", command.flags.cachePairs, DEFAULT_CACHE_PAIRS)
		: undefined;
	const cacheConcurrency = cacheMode
		? normalizePositiveInteger("cache-concurrency", command.flags.cacheConcurrency, DEFAULT_CACHE_CONCURRENCY)
		: undefined;
	const runs = cacheMode
		? cachePairs! * 2
		: normalizePositiveInteger("runs", command.flags.runs, PROFILE_DEFAULT_RUNS[profile]);
	const maxTokensOverride =
		command.flags.maxTokens !== undefined
			? normalizePositiveInteger("max-tokens", command.flags.maxTokens, 1)
			: undefined;

View on GitHub (pinned to 9690622007)

Solutions

1. Add --profile chat or --profile generation when passing --prompt
2. Remove --prompt if you want the default mix workload
3. Use --prefill-bytes with --profile prefill instead if you intend a prefill benchmark

Example fix

// before
omp bench sonnet --prompt "Explain quines"
// after
omp bench sonnet --prompt "Explain quines" --profile chat

if (args.prompt !== undefined && !(args.profile === "chat" || args.profile === "generation"))
  throw new Error("--prompt requires --profile chat or generation");

try {
  await runBench(argv);
} catch (err) {
  if (err instanceof Error && err.message.includes("--prompt requires")) {
    console.error("Add --profile chat/generation, or drop --prompt.");
    process.exitCode = 2;
  } else throw err;
}

When it happens

Trigger: Running e.g. `omp bench sonnet --prompt "hi" --profile mix` or `--prompt "hi" --profile prefill` (with --cache absent).

Common situations: Passing --prompt with the default profile (mix) without realizing the default isn't chat; scripts that always include --prompt.

Related errors

• --prefill-bytes requires prefill challenges (--profile mix o
• Use --cache-pairs instead of --runs with --cache
• --cache builds its own stable-prefix prompts
• --profile cannot be combined with --cache
• --par cannot parallelize cold/warm pairs; use --cache-concur

AI-assisted analysis of can1357/oh-my-pi@9690622007 (2026-08-31). Data as JSON: /api/errors/7d0eb356cdfaada2. Report an issue: GitHub.