GroqCloud vs Vertex AI (Gemini Enterprise Agent Platform)

GroqCloud 64

No OpenAPI document in the docs, and the SDK's .stats.yml carries only an endpoint count of 17, no spec URL (0). llms.txt at console.groq.com/llms.txt (10). Reference not read in full this run (15 of 20). Structured outputs have a Strict mode as well as Best-effort (15). The errors page lists 15 status codes with recovery advice, including 498 for Flex capacity and 424 for remote MCP auth, and the error object with message and type (14 of 15). The changelog at console.groq.com/docs/changelog, read on 1 November, is dated, but its newest entry is 18 April 2026 and it doesn't record the four shutdowns since July, which only the deprecations page carries (10 of 15).

Vertex AI (Gemini Enterprise Agent Platform) 94

Model reading. Keyless discovery document for the Agent Platform API v1, revision 20260930, with 1,128 methods and 1,493 schemas (25). No llms.txt, but every docs page answers as Markdown at .md.txt, counted as for vertex-ai-tuning (10). Model cards give limits, capabilities and retirement dates, and the 3.8 Flash card says when to use 3.7 Flash instead, but the docs mix the Vertex AI and Agent Platform names, and the OpenAI page tells new users to call the Gemini Developer API instead (16 of 20). Structured output with response_schema and response_mime_type: application/json always follows the schema per the docs, and complex schemas fail with a 400 (15). Examples on every page and a 429 guide, but we found no full error code list for generateContent (13 of 15). Versioned v1 and v1beta1 APIs and release notes with entries on most working days, newest 5 November 2026 (15).

GroqCloud 72

Model reading. Meta's Muse Spark Open went into limited preview on GroqCloud on 15 October 2026, per Groq's announcement, 17 days before this check, and the models page lists it as an enterprise model (30). The newest shutdown is still groq/compound and compound-mini on 21 September. The changelog's newest entry is 18 April 2026, the Python SDK 1.7.0 (tagged 25 August) and TypeScript SDK 1.6.0 are still the newest tags, with commits to 4 September and no release since, and the deprecations page has no announcement after 24 August. Production models get an email and a migration path, previews may go at short notice, and no minimum period is stated. Llama 3.1 8B and 3.3 70B got 60 days on the free and developer tiers, Compound 28 days with no replacement named (4 of 12). Three shutdown dates in the 90 days to 1 November, 16 August, 14 September and 21 September. Three in 90 days is still frequent churn (0 of 8). A dated changelog with an RSS feed and a community forum exist, but the changelog's newest entry is 18 April 2026; support replies not checked (10 of 15). SDK issue replies not sampled, since GitHub's API refused our shell (5 of 10). Python SDK 1.7.0 and TypeScript SDK 1.6.0, both published on 26 August 2026 and still the newest tags on 1 November (15). CI, release-please and lock-file vulnerability fixes in 1.7.0 (8 of 10). 30+4+0+10+5+15+8 = 72.

Vertex AI (Gemini Enterprise Agent Platform) 91

Model reading. Gemini Omni 1.2 Flash GA on 5 November, Grok 4.7 GA on 27 October, Gemini 4 Argon GA on 13 October and Claude Haiku 5.5 on 7 October 2026 (30). Stable models stay at least 12 months, and since 6 October Gemini 4 Argon and Gemini 3.8 Flash are in that tier, while 3.7 and 3.6 Flash remain short-term with at least 45 days' notice once a date is set (10 of 12, up from 8). Gemini 2.5 Pro, Flash and Flash-Lite were due to retire on 20 October, three models on one date after more than 12 months, with replacements named. We didn't check partner model retirements (6 of 8, down from 8). Dated release notes on most working days, newest 5 November, and Cloud Customer Care (15). python-genai replies not sampled this run (5 of 10). Current SDKs, google-genai 3.0.0 on 6 November and @google/genai 3.0.0, and google-cloud-aiplatform 2.7.0 on 28 October 2026 (15). Supported runtimes stated, Python 3.10 or later and Node 20 or later, per the 6 October check (10). Unchanged at 91.

GroqCloud 80

Model reading of the checklist (tool use, structured output, caching, context, batch, SDKs, errors). Tool use documented, parallel calls and forced choice not checked (15 of 20). Strict structured outputs (15). Automatic prompt caching at 50% off cached input, with a 2-hour cache life, on gpt-oss-20b, gpt-oss-120b and gpt-oss-safeguard-20b only (10 of 15). 131,072-token context on every self-serve model; only MiniMax M2.7, a preview on enterprise pricing, reaches 196,608 (5 of 15). Batch at half price (10). Official SDKs for Python and JavaScript (10). Errors page with codes, recovery advice and a typed error object (15).

Vertex AI (Gemini Enterprise Agent Platform) 97

Model reading of the checklist (tool use, structured output, caching, context, batch, SDKs, errors). Function calling with FunctionCallingConfig mode ANY and allowedFunctionNames to force a named tool (20). Schema-guaranteed JSON output (15). Implicit and explicit context caching, cached input at 0.1x, $0.075 per million on 3.8 Flash through 31 December (15). 1,048,576-token context on 3.8 Flash, a unified 1,048,576-token input and output limit on Gemini 4 Argon, and 1,000,000 on Claude Opus 5.5 (15). Flex/Batch at half price on Gemini (10). Gen AI SDKs for Python and JavaScript among others, and the OpenAI libraries work with Google Cloud auth (10). 429 handling documented per quota framework, with standard Google status codes elsewhere (12 of 15).

GroqCloud 82

Model reading. Bearer keys scoped to a project, with custom request limits per project and model permissions at organisation and project level. The security onboarding page advises rotating keys periodically and revoking them at once if compromise is suspected (30). The services agreement, read first-hand on 6 October and not re-read on 17 October, says Groq is not permitted to use inputs or outputs for training unless the customer grants permission (20). No retention by default, up to 30 days for reliability and abuse monitoring, and zero retention as a setting any customer can turn on in Data Controls, per the data page (15). Request logs, usage per project and a read-only Reader role (12 of 15). groq.com's security.txt holds a Contact line and nothing else, and the trust centre at trust.groq.com is a Vanta page that renders only with JavaScript, so certifications and any bug bounty are unchecked (5 of 20).

Vertex AI (Gemini Enterprise Agent Platform) 97

Model reading. OAuth 2.0 with IAM roles and service accounts, which the docs recommend for production, with API keys for testing. No key in a query string found in the docs we read (30). The training restriction in the Service Specific Terms says Google won't use customer data to train or fine-tune any model without permission, GA and pre-GA alike (20). Default in-memory caching for 24 hours that can be disabled, flagged prompts logged up to 90 days unless an abuse-monitoring exception is granted or the customer has a Cloud Master Agreement, the Interactions API storing by default, and Claude Fable 5 and Mythos 5 requiring 30-day retention shared with Anthropic (12 of 15, a judgement call matching the Fable exception on the Anthropic and Bedrock listings). Cloud Audit Logs, with Data Access logs off until enabled, and optional request-response logging to BigQuery (15). google.com security.txt valid to 2030-04-01, the Google VRP, and Generative AI on the platform in ISO 27001, 27017 and 27018 and SOC 1, 2 and 3 scope per the 1 October services-in-scope check for vertex-ai-tuning, not re-read on 6 October (20).

GroqCloud 90

Status page at groqstatus.com on incident.io, with components for the API, each production and preview model and the website (20). The page's RSS feed and the incident pages it links, read on 1 November, list three incidents in the 90-day window, all in October, and nothing since 12 October apart from a console maintenance on 18 October. On 10 October the enterprise-only preview MiniMax M2.7 had failed requests from data centre hardware issues from 10:09 to 16:14 UTC. The incident's header and impact record say degraded performance, while its Identified update set the model to partial outage and warned of failed requests. On 12 October console.groq.com returned 500 errors on metrics, /models and key creation from 03:17 to 03:38, and the vendor said inference requests were not affected. On 12 October MiniMax M2.7 had a partial outage from 18:17 to 18:36. None took down a core API for an hour, since the six-hour incident hit one enterprise preview model that the status page describes as for evaluation only, so minor incidents only (20). A stricter reading would count it as a major and give 10. Limits published per model, 30 requests a minute, 1,000 a day and 8,000 tokens a minute on gpt-oss for the free plan, re-read on 1 November (15). 429 carries retry-after, x-ratelimit-* headers come on every response, and the errors page says to wait before retrying a 503 and that a 502 may clear on retry (15). The Performance Tier, enterprise only, lists a 99.9% availability SLA with details in the offline agreement, per the 6 October check, not re-read on 1 November (10). gpt-oss and Whisper models are production, Qwen 3.8 27B, MiniMax M2.7 and Muse Spark Open are previews (10). 20+20+15+15+10+10 = 90.

Vertex AI (Gemini Enterprise Agent Platform) 99

Google Cloud status dashboard with per-product history (20). Its incident feed, read on 6 November, has no incident naming Vertex Gemini API, Gemini or the Agent Platform in the 90 days. The regional incidents of 20 August, 1 September and 20 October (europe-west2) and the Cloud Storage incident of 8 and 9 October didn't list it among affected products (30). Standard PayGo publishes usage tiers by 30-day spend with baseline throughput, 2,000,000 to 50,000,000 tokens a minute for Flash models and 500,000 to 10,000,000 for Pro, plus quota tables for other methods (15). A 429 page says to use the global endpoint and truncated exponential backoff, and a retry-strategy page (15). SLA of 99.5 per cent a month on generateContent and streamGenerateContent, 95 per cent for short-term availability models, and a latency SLO on Provisioned Throughput. The 429 page says pay-as-you-go 429s don't count toward the SLA error rate (10). Gemini 4 Argon reached GA on 13 October and Gemini 3.8 Flash moved to the 12-month tier, and the v1 API, 3.5 Flash-Lite and Claude Opus 5.5 are GA, but express mode was Preview at the 6 October check (9 of 10, up from 8). Up from 98 on 6 October.

GroqCloud 85

Closed service with clear terms, SDKs Apache-2.0 (15). The data page agrees with the listing on retention and adds detail, batch files kept 30 days unless deleted, fine-tuning data kept until the customer deletes it, and SCCs for transfers. The services agreement bars training on inputs and outputs, and the DPA gives 15 days' notice of sub-processor changes to subscribers, both read on 6 October. The agreement extends zero retention to 'Eligible Customers' where the data page says all customers may enable it (26 of 30). Deprecations page with announcement and shutdown dates (20). All customer data in Google Cloud buckets in the US, per the data page. The DPA points to a sub-processor list at trust.groq.com/subprocessors, which needs JavaScript and wasn't read (10 of 20).

Vertex AI (Gemini Enterprise Agent Platform) 90

Closed service under the Google Cloud terms, with the Gen AI SDKs under Apache-2.0 and a public discovery document (18, as for Bedrock's open service models). The zero data retention and abuse monitoring pages agree with the Cloud terms and give periods for each case, 24 hours in memory, up to 90 days for flagged prompts, 30 days for Advanced AI models, 3 days for Search grounding logs and 30 days for Maps grounding, and say which can't be switched off (27 of 30). A dated model versions table with release and retirement dates and a stated 12-month or 45-day policy, but on 6 November it still listed Gemini 2.5 Pro, Flash and Flash-Lite as available with a 20 October 2026 retirement date and left them out of the retired table (18 of 20, down from 20). The Cloud subprocessor list with locations, per the 1 October vertex-ai-tuning check, regional and multi-region endpoints, and per-model ML processing locations on the partner model cards (20).