test: hold the Groq example count at two so the daily quota fits

Serialising by provider stopped Groq examples competing for the same quota window,
but the binding limit is a daily one: the org is capped at 100k tokens per day on
this tier, and four Groq examples making several calls each do not fit inside it.
No retry policy or backoff addresses that.

So the two carrying the least distinct information come out:

  groq-code-interpreter-js       - the same demo as its Python twin, which stays
  upload-dataset-code-interpreter - a third Groq chart demo

groq-code-interpreter-python and mcp-groq-exa-js remain, which still covers both
Groq SDK surfaces. Roughly half the token spend for nearly all the coverage.

Both dropped examples work - the exclusion note says so explicitly, because
"excluded" has meant "broken" for every other entry in that list and these two
should go straight back if the tier is raised. They are not being quietly retired.

Also fixed a stale reference in the runner's header comment: it still described
results.json as feeding updateTestsMd.js, which report.mjs replaced.

Author: beran-t

tests/run-examples.ts

@@ -8,8 +8,8 @@
  * e2b SDK: e2b depends on chalk 5 and four other ESM-only packages, and while
  * Node 22+ resolves those fine from both CJS and ESM, Jest's own runtime does
  * not implement `require(esm)`. Rather than fight that, the runner is a plain
- * script. It still writes tests/results.json in the shape updateTestsMd.js
- * expects, so the reporting and Slack steps are unchanged.
+ * script. It still writes tests/results.json, which tests/report.mjs turns into
+ * the artifact table and the Slack payload.
  */
 import { promises as fs } from 'fs'
 import path from 'path'
@@ -27,12 +27,11 @@ const scripts: {
   interpreter: Interpreter
   file: string
   // Providers that rate-limit per organisation. Examples sharing a provider are
-  // serialised against each other, which fixed one of the three Groq failures.
-  // The residual limit is not concurrency though: the Groq org is capped at
-  // 100k tokens per DAY on this tier ("tokens per day (TPD): Limit 100000,
-  // Used 99952 ... try again in 10m"), and four Groq examples several calls each
-  // can exhaust it - especially across repeated runs on one day. No retry policy
-  // fixes a daily cap; it needs a higher tier or fewer Groq examples per run.
+  // serialised against each other, so they never compete for the same quota
+  // window. That is not enough on its own: the Groq org is capped at 100k tokens
+  // per DAY on this tier ("tokens per day (TPD): Limit 100000, Used 99952"), which
+  // no retry policy can work around, so the Groq example count is deliberately
+  // held at two. See the exclusion note below before adding a third.
   provider?: string
   // What to pass after `uv run` / `poetry run`. Defaults to main.py for uv and
   // the `start` console script for poetry. Can be a path, a script name, or
@@ -48,14 +47,12 @@ const scripts: {
   { name: 'codestral-code-interpreter-js', interpreter: 'npm', file: './examples/codestral-code-interpreter-js/' },
   { name: 'gpt-4o-code-interpreter-js', interpreter: 'npm', file: './examples/gpt-4o-js/' },
   { name: 'codestral-code-interpreter-python', interpreter: 'jupyter', file: './examples/codestral-code-interpreter-python/codestral_code_interpreter.ipynb' },
-  { name: 'upload-dataset-code-interpreter', provider: 'groq', interpreter: 'jupyter', file: './examples/upload-dataset-code-interpreter/llama_3_code_interpreter_upload_dataset.ipynb' },
   { name: 'hello-world-python', interpreter: 'poetry', file: './examples/hello-world-python/' },
   { name: 'o1-code-interpreter-js', interpreter: 'npm', file: './examples/o1-and-gpt-4-js/' },
   { name: 'gpt-4o-code-interpreter', interpreter: 'jupyter', file: './examples/gpt-4o-python/gpt_4o.ipynb' },
   { name: 'together-ai-code-interpreter-python', interpreter: 'jupyter', file: './examples/together-ai-code-interpreter-python/together_with_e2b_code_interpreter.ipynb' },
   { name: 'langchain-python', interpreter: 'poetry', file: './examples/langchain-python/' },
   { name: 'langgraph-python', interpreter: 'poetry', file: './examples/langgraph-python/' },
-  { name: 'groq-code-interpreter-js', provider: 'groq', interpreter: 'npm', file: './examples/groq-code-interpreter-js/' },
   { name: 'claude-code-interpreter-python', interpreter: 'jupyter', file: './examples/claude-code-interpreter-python/claude_code_interpreter.ipynb' },
   { name: 'claude-visualize-website-topics', interpreter: 'jupyter', file: './examples/claude-visualize-website-topics/claude-visualize-website-topics.ipynb' },
   { name: 'mcp-client-js', interpreter: 'npm', file: './examples/mcp-client-js/' },
@@ -98,6 +95,15 @@ const scripts: {
 // against the current CLI - either pin Codex in the template or wire its current
 // auth - so it is not guessed at here:
 //   openai-codex-in-sandbox-js, openai-codex-in-sandbox-python
+// Dropped to fit the Groq quota, NOT because they are broken - both pass when the
+// day's budget allows, and they should be restored if the tier is raised. The org
+// is capped at 100k tokens per day and four Groq examples several calls each do not
+// fit, so the two carrying the least distinct information come out:
+// groq-code-interpreter-js is the same demo as its Python twin, which stays, and
+// upload-dataset-code-interpreter is a third Groq chart demo.
+// groq-code-interpreter-python and mcp-groq-exa-js remain, covering the two
+// different SDK surfaces:
+//   groq-code-interpreter-js, upload-dataset-code-interpreter
 // Calls a model the Fireworks account cannot reach: qwen2p5-coder-32b-instruct
 // returns 404 "Model not found, inaccessible, and/or not deployed", which does not
 // distinguish a retired model from one this account has not deployed. Needs someone