gaslamp 2.1.1 → 2.1.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/package.json +1 -1
- package/src/consult.mjs +5 -1
- package/src/setup.mjs +11 -6
package/package.json
CHANGED
package/src/consult.mjs
CHANGED
|
@@ -386,7 +386,11 @@ export function runConsult(backend, argv) {
|
|
|
386
386
|
costUsd: j.total_cost_usd ?? null,
|
|
387
387
|
};
|
|
388
388
|
} else if (backend === "codex" && j.type === "turn.completed" && j.usage) {
|
|
389
|
-
|
|
389
|
+
// OpenAI semantics differ from Anthropic's: input_tokens already
|
|
390
|
+
// INCLUDES the cached portion (cached_input_tokens is a subset, and
|
|
391
|
+
// codex's own total_tokens = input + output), so adding it again
|
|
392
|
+
// double-counts cache hits.
|
|
393
|
+
usageAcc.input += j.usage.input_tokens ?? 0;
|
|
390
394
|
usageAcc.output += j.usage.output_tokens ?? 0;
|
|
391
395
|
meta.usage = { inputTokens: usageAcc.input, outputTokens: usageAcc.output, costUsd: null };
|
|
392
396
|
} else if (backend === "codex" && j.type === "item.completed") {
|
package/src/setup.mjs
CHANGED
|
@@ -91,25 +91,30 @@ Fleets default to \`--sandbox read-only\` (N writers in one cwd race) — pass \
|
|
|
91
91
|
|
|
92
92
|
## Picking a model
|
|
93
93
|
|
|
94
|
+
Both ladders, strongest first: fable ≈ astra > opus ≈ sol > terra > sonnet ≈ luna > haiku.
|
|
95
|
+
|
|
94
96
|
${other === "codex"
|
|
95
|
-
? `Codex tiers
|
|
97
|
+
? `Codex tiers — pass the full id to \`--model\` (bare tier names are rejected):
|
|
96
98
|
|
|
97
99
|
| tier | \`--model\` | ≈ claude | reach for it when |
|
|
98
100
|
|------|-----------|----------|-------------------|
|
|
99
|
-
|
|
|
100
|
-
|
|
|
101
|
+
| astra | \`gpt-6-astra\` | fable | the hardest reasoning: adversarial verification, design spars, proofs |
|
|
102
|
+
| sol | \`gpt-5.6-sol\` | opus | substantial work: reviews, diagnosis, real implementation |
|
|
103
|
+
| terra | \`gpt-5.6-terra\` | (between opus and sonnet) | everyday work: routine reviews, well-scoped changes |
|
|
101
104
|
| luna | \`gpt-5.6-luna\` | sonnet | quick checks, mechanical transforms, high-N fleets |`
|
|
102
|
-
: `Claude tiers
|
|
105
|
+
: `Claude tiers — \`--model\` takes the bare alias:
|
|
103
106
|
|
|
104
107
|
| \`--model\` | ≈ codex | reach for it when |
|
|
105
108
|
|-----------|---------|-------------------|
|
|
106
|
-
| \`fable\` |
|
|
107
|
-
| \`opus\` |
|
|
109
|
+
| \`fable\` | astra | the hardest reasoning: adversarial verification, design spars, proofs |
|
|
110
|
+
| \`opus\` | sol | substantial work: reviews, diagnosis, real implementation |
|
|
108
111
|
| \`sonnet\` | luna | quick checks, mechanical transforms, high-N fleets |
|
|
109
112
|
| \`haiku\` | (below luna) | pings, one-word sanity checks, the cheapest fan-outs |`}
|
|
110
113
|
|
|
111
114
|
Omit \`--model\` to use ${Other}'s own configured default. Match the tier to the stakes: verification you'll rely on deserves the top tier; a fleet vote usually doesn't.
|
|
112
115
|
|
|
116
|
+
\`--effort\` takes \`low\` / \`medium\` / \`high\` / \`xhigh\` / \`max\`${other === "codex" ? "; astra, sol and terra also take `ultra` (max reasoning plus automatic task delegation). Omitted = Codex's configured `model_reasoning_effort`, else the model's own default — sol's is `low`, so pass it explicitly when the consult matters." : ". Omitted = Claude's own default."}
|
|
117
|
+
|
|
113
118
|
## Reading the records
|
|
114
119
|
|
|
115
120
|
gaslamp jobs [--json] # recent consults: status, label, thread, tokens
|