summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorIgor Soarez <igor@soarez.org>2026-11-20 10:54:44 +0000
committerIgor Soarez <igor@soarez.org>2026-11-20 10:54:44 +0000
commitfcf5c24b66a013672a6d9dcac7ab09b813b4f5c7 (patch)
tree4c542613b56a45266f1d1a48c274b024a737f11a
parentdcbd341a04bbbf0ac7d720a3a3bf882778983bfe (diff)
pi-review-panel, pi-subagent: GLM through zai, the z.ai GLM Coding Plan, not per-token providers
-rw-r--r--skills/pi-review-panel/SKILL.md14
-rw-r--r--skills/pi-subagent/SKILL.md10
2 files changed, 19 insertions, 5 deletions
diff --git a/skills/pi-review-panel/SKILL.md b/skills/pi-review-panel/SKILL.md
index f9712f9..22721a9 100644
--- a/skills/pi-review-panel/SKILL.md
+++ b/skills/pi-review-panel/SKILL.md
@@ -29,10 +29,20 @@ The default panel is six runs:
|---|---|---|
| fireworks | `accounts/fireworks/models/kimi-k3` | fw-kimi-k3 |
| openrouter | `moonshotai/kimi-k3` | or-kimi-k3 |
-| openrouter | `z-ai/glm-5.3-flash` | or-glm-5.3-flash |
+| zai | `glm-5.3-flash` | zai-glm-5.3-flash |
| openrouter | `deepseek/deepseek-v4.1-flash` | or-deepseek-v4.1-flash |
| openrouter | `google/gemini-3.7-flash` | or-gemini-3.7-flash |
-| openrouter | `z-ai/glm-5.3` | or-glm-5.3 |
+| zai | `glm-5.3` | zai-glm-5.3 |
+
+**GLM runs on the z.ai GLM Coding Plan** (provider `zai`), a flat monthly
+subscription, not OpenRouter or Fireworks, which bill per token. A GLM model the
+user names without a provider ("glm 5.3", "glm flash") means `zai`. The plan has
+a shared allowance (2,000 credits per 5 hours, 10,000 a week, across every pi
+session on every host); a run that hits it fails with a quota error. Report
+that run as failed for that reason; do not rerun it on OpenRouter or Fireworks
+unless the user says so. pi still prices zai runs at z.ai's pay-as-you-go
+rates, so in the cost table mark those rows "plan (not billed)" and leave them
+out of the total.
If the user names specific providers/models, use exactly those instead — don't
silently add the defaults back. A provider this machine lacks
diff --git a/skills/pi-subagent/SKILL.md b/skills/pi-subagent/SKILL.md
index 9245b0e..d2ac667 100644
--- a/skills/pi-subagent/SKILL.md
+++ b/skills/pi-subagent/SKILL.md
@@ -43,7 +43,11 @@ output, it's a sequence, not a split — run it in stages.
others. If they named none, ask — or, for a fan-out, offer the default panel in
the pi-review-panel skill. Check a name resolves first:
`pi --offline --list-models <search>`. Credentials are already configured in
-`pi`.
+`pi`. GLM models go through `zai`, the z.ai GLM Coding Plan (flat monthly,
+with a shared 5-hour and weekly allowance), rather than OpenRouter or Fireworks;
+pi's cost for a zai run is the pay-as-you-go price, not what is billed, so
+report it as covered by the plan. A run that hits the plan's limit fails: say
+so rather than rerunning it on a pay-per-token provider unasked.
## 1. Set up a workdir
@@ -128,7 +132,7 @@ A fan-out of one brief to three models:
```bash
"$SK/run-agent.sh" "$P" fireworks "accounts/fireworks/models/kimi-k3" fw-kimi-k3 > "$P/out/_fw-kimi-k3.log" 2>&1 &
-"$SK/run-agent.sh" "$P" openrouter "z-ai/glm-5.3" or-glm-5.3 > "$P/out/_or-glm-5.3.log" 2>&1 &
+"$SK/run-agent.sh" "$P" zai "glm-5.3" zai-glm-5.3 > "$P/out/_zai-glm-5.3.log" 2>&1 &
"$SK/run-agent.sh" "$P" openrouter "google/gemini-3.7-flash" or-gemini > "$P/out/_or-gemini.log" 2>&1 &
```
@@ -136,7 +140,7 @@ A split, each part with its own brief:
```bash
for part in parser cli docs; do
- "$SK/run-agent.sh" --brief "$P/brief-$part.md" "$P" openrouter "z-ai/glm-5.3" "$part" \
+ "$SK/run-agent.sh" --brief "$P/brief-$part.md" "$P" zai "glm-5.3" "$part" \
> "$P/out/_$part.log" 2>&1 &
done
```