Head to head · Fine-tuning · October 2026 research run

Microsoft Foundry fine-tuning (Azure OpenAI) vs Tinker

Microsoft Foundry fine-tuning (Azure OpenAI) scores 59.2 (C) on agent readiness against Tinker's 53.7 (D), and leads in 4 of 7 scored categories. Tinker leads on agent ergonomics and maintenance & community. Both do fine-tuning.

Best fine-tuning services for AI models · All 36 fine-tuning comparisons

Which one, for what

Microsoft Foundry fine-tuning (Azure OpenAI) C

Best for Teams that must tune an OpenAI model, need Azure's compliance and regional controls, and will serve the result on Azure.

Ahead on

  • Reliability, 77 against 35
  • Schema & documentation, 77 against 70
  • Security & auth, 85 against 57
  • Transparency & trust, 82 against 62

Also in its favour

  • A hosted endpoint, with nothing to install

Watch for

No weight export; checkpoints copy only between Azure resources

Tinker D

Best for Researchers and teams writing custom post-training loops, especially RL, who want per-token billing and the weights at the end.

Ahead on

  • Agent ergonomics, 53 against 47
  • Maintenance & community, 87 against 55

Also in its favour

  • Open source
  • No incidents deducted, where Microsoft Foundry fine-tuning (Azure OpenAI) loses 7 points for them

Watch for

LoRA only; no full-parameter training

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reliability Microsoft Foundry fine-tuning (Azure OpenAI) +42

Microsoft Foundry fine-tuning (Azure OpenAI) 77

Hosted reading. Azure status page with post-incident reviews kept for five years, Azure OpenAI Service and Foundry Models among its products (20). In the 90 days from 4 August the history holds two reviews and none after 30 September. The 29 September 2026 review covers intermittent failures and latency for Azure OpenAI, Foundry Models and Azure AI services in Sweden Central from 10:03 to 15:58 UTC, one region and a share of requests, so partial and minor, as the Azure Speech, Translator, Document Intelligence and Key Vault listings read it. The 30 September to 1 October gateway review names networking services only (20, was 10). Fine-tuning limits published with numbers, 3 simultaneous training jobs (5 on the developer tier), 20 queued, 100 jobs and 100 files per resource, 2 billion tokens per job and 720 hours (15). The quota page says to retry with backoff and links code samples, with no safe-retry guidance for job creation (10). The November 2026 Online Services SLA gives Foundry Models 99.9 per cent with 10 and 25 per cent credits, measured on requests to model endpoints, so it covers serving a tuned model but not the fine-tuning job API, and excludes the Developer tier. Read as the Teams and Outlook mail levels, which measure the service's function but not the API (7 of 10). Standard and global training are documented without a preview label, while Programmed Training, the developer tier and seven of 18 tunable models are preview and GPT-5 RFT is by invitation (5). 20 + 20 + 15 + 10 + 7 + 5 = 77.

Tinker 35

The listing's kind is sdk, but the SDK is a thin client for a hosted training service, so the hosted checklist is kept as on 1 October; that's a judgement call. No status page found, on the vendor's pages read on 1 October or by search on 4 November (0), and so no readable incident history (5). The only published limit is the in-flight sample cap of 2,000, per the 1 October check, not re-read on 4 November (10). The SDK's retry handler retries sampling with stable request IDs, confirmed in the SDK repository, and the docs' guide on client-side retries is per the 1 October check (15). No SLA found (0). Sign-up is open and paid, but the cookbook README still says 'after our private beta is over' on 4 November, and no GA statement was found (5).

Maintenance & community Tinker +32

Microsoft Foundry fine-tuning (Azure OpenAI) 55

Read as a closed service with official SDKs, from the vendor's dated sources as the other Azure listings are. The newest dated change that touches fine-tuning is the fine-tuning sessions API (preview, create, list, forward and backward passes, optimiser steps, checkpoints and sampling under /fine_tuning/sessions), added to Microsoft's public REST specification in the Foundry release of 30 October 2026, 3 days back (30). The specification doesn't name Programmed Training, but its training-step calls match the Programmed Training preview the docs describe. In the 90 days from 4 August that is the only dated service change touching fine-tuning. The older fine-tuning specification files changed only on 13 August (a lint suppression), the Foundry what's new page (dated 1 September) covers August by month with no fine-tuning entry, the Azure OpenAI what's new page's newest section is May 2026 and the Azure Developer CLI fine-tuning extension's newest release is 0.0.17-preview (20 February 2026). azure-ai-projects 2.8.0 (2 October), the Foundry SDK the Python how-to now uses, adds supervised and reinforcement fine-tuning data-generation configurations and counts half. With the SDK release at half that counts 1.5, short of three (0, was 10). Microsoft Q&A and paid support exist, and the fine-tuning pages are updated (29 October) though no fine-tuning changelog exists (10). The openai SDKs the docs point to are maintained by OpenAI, and Microsoft's azure-ai-projects is current (10). Package health not checked (5). 30 + 0 + 10 + 10 + 5 = 55.

Tinker 87

tinker-cookbook v0.5.9 tagged on 27 October 2026, 8 days before this check. The SDK repository's newest version bump is 0.37.0 on 2 November, untagged, so it isn't counted (30). Cookbook tags v0.5.4 (11 August), v0.5.5 (19 August), v0.5.6 (1 September), v0.5.7 (3 September), v0.5.8 (21 October) and v0.5.9 (27 October) in the 90 days from 6 August (20). The cookbook had 22 open issues and 55 open pull requests on 1 October and its README defers outside PRs until the private beta ends; GitHub's pages refuse us, so not re-read (12). The SDK is current (15). The cookbook runs pytest on push and pull request, plus pyright, recipe smoke tests and a security workflow (10).

Security & auth Microsoft Foundry fine-tuning (Azure OpenAI) +28

Microsoft Foundry fine-tuning (Azure OpenAI) 85

Microsoft Entra ID tokens with Azure RBAC, or two rotatable resource keys in the api-key header; the key gives full data-plane access to the resource (30). Training needs Foundry User and deploying needs Foundry Owner, so the roles split the two (15). Returns job state and your own model's output, no third-party content (10). Azure Monitor resource logs, once a diagnostic setting is created, and the subscription activity log, for every customer (15). Microsoft's coordinated vulnerability disclosure and cloud bounty programmes (up to $100,000) and a SOC 2 Type 2 attestation for Azure; microsoft.com's security.txt passed its Expires date on 2026-09-23 and still carries it on 2 November. MSRC publishes CVEs for Azure AI Foundry, counted under negative events (15). 30 + 15 + 10 + 15 + 15 = 85.

Tinker 57

API keys from the console, revocable, with key verification since 0.26.2; no per-key scopes found (20). An org, team and project permissions model, with no read-only role found (5). Returns your own model's samples and losses, no third-party content (10). get_audit_log() in the SDK for admins (15). security.txt at thinkingmachines.ai lists security-reports@thinkingmachines.ai and expires 2029-07-13, per the 2 November reading this run made, and the site now links a disclosure policy that wasn't read. A reporting route with no stated acknowledgement time, up from 5 (7). No bug bounty, certifications or advisories found, and osv.dev lists none for tinker or tinker-cookbook.

Transparency & trust Microsoft Foundry fine-tuning (Azure OpenAI) +20

Microsoft Foundry fine-tuning (Azure OpenAI) 82

Closed service under the Microsoft Product Terms, whose generative AI clause rules out training foundation models on Customer Data, and the openai SDK is Apache-2.0. Closed with clear terms is 15 on the checklist, as the other Azure listings read it (15, was 20). The data privacy page (updated 5 June 2026, re-read on 2 November) agrees with the terms. Training files and tuned models stay in the resource's geography, are encrypted and can be deleted at any time and are exclusive to the customer, and training data isn't used to train foundation models. Abuse-monitoring retention for fine-tuning isn't stated (26). A written retirement policy of at least 18 months after GA and 60 days' notice, with separate training and deployment retirement dates for each tunable OpenAI model (October 2027 for deployments) (20). Data locations are stated per training and deployment type, including that global training may process in any geography; the subprocessor list wasn't checked (15). 15 + 26 + 20 + 15 = 76.

Tinker 62

The SDK and the cookbook are Apache-2.0, and the hosted service has Service Terms of Use, last updated 10 January 2026, at thinkingmachines.ai/legal/terms/, which the 1 October check didn't find. Closed service with clear terms (15, up from 5). The terms commit not to use Customer Content for model training or fine-tuning, keep it only as needed for the service, and carry a DPA with deletion on request. Checkpoints carry user-set TTLs. The privacy notice moved to /legal/privacy/ and wasn't re-read (15, up from 8). A model deprecations page lists dated retirements and promises to 'aim to give advance notice' by email, per the 1 October check. Dated notices with no stated notice period, up from 12 (15). The DPA's Annex 2 names Google Cloud, Cloudflare, WorkOS and OpenAI with their service and location, mostly 'Global' (15, up from 5). 15 + 15 + 15 + 15 = 60.

Schema & documentation Microsoft Foundry fine-tuning (Azure OpenAI) +7

Microsoft Foundry fine-tuning (Azure OpenAI) 77

A public OpenAPI 3.2 document for the v1 data plane, "Azure AI Foundry Models Service", in Azure/azure-rest-api-specs (specification/ai/data-plane/OpenAI.v1), with 12 fine-tuning paths among 71, from jobs, events, checkpoints and copy to pause, resume and graders (25). learn.microsoft.com has no llms.txt, but its pages answer Accept: text/markdown with Markdown, as the other Azure listings read it (5). The how-to guides say when to use SFT, DPO or RFT and list which models take which (15). The job body is the OpenAI shape, typed by method with hyperparameters per method (12). Python, REST, portal and Azure Developer CLI examples. No fine-tuning error reference found (8). Versioned by api-version and /openai/v1, but the Azure OpenAI what's new page's newest section is still May 2026 (12). 25 + 5 + 15 + 12 + 8 + 12 = 77.

Tinker 70

No REST reference or OpenAPI. The contract is the typed Python SDK, whose API reference ships in the SDK repository's docs folder (15). llms.txt, llms-full.txt, models.json and serverless.json per the 1 October check, not re-read on 4 November, since the docs host is paused on its terms (10). Reference pages for ServiceClient, TrainingClient, SamplingClient and RestClient and an SDK cheatsheet (12). Typed parameter objects and bounded checkpoint TTLs (10). The cookbook's recipes are the examples; no page of error types found (8). Semver 0.x releases with a dated changelog that names removals, per the 1 October check (15).

Agent ergonomics Tinker +6

Microsoft Foundry fine-tuning (Azure OpenAI) 47

Job objects are compact and list calls take limit; no field selection (10). Cursor pagination with after on the OpenAI-shaped lists; no filters found (10). Errors come back in the OpenAI error shape with a code and message; no fine-tuning error table (12). No idempotency key on job creation (5). Only model and training_file are required and the openai SDK exists in several languages, but deploying the result needs a second call to the Resource Manager API under a different role (10).

Tinker 53

Sampling takes token limits and logprob options, so responses can be sized; no field selection (15). RestClient filters training runs by project_id since 0.22.6, and the console filters sessions by status (10). No documented error codes found (5). Stable request IDs make SDK retries of sampling safe; nothing on retrying an optim_step (15). A LoRA client needs only a base model and a rank, but the SDK is Python only (8).

Score by category

CategoryWeightMicrosoft Foundry fine-tuning (Azure OpenAI)TinkerEdge
Reliability16%7735Microsoft Foundry fine-tuning (Azure OpenAI) +42
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%7770Microsoft Foundry fine-tuning (Azure OpenAI) +7
Agent ergonomics13%4753Tinker +6
Security & auth14%8557Microsoft Foundry fine-tuning (Azure OpenAI) +28
Payments & pricing10%3030even
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%5587Tinker +32
Transparency & trust7%8262Microsoft Foundry fine-tuning (Azure OpenAI) +20
Negative events≤15-70
TotalC 59.2/100D 53.7/100

Facts side by side

FactMicrosoft Foundry fine-tuning (Azure OpenAI)Tinker
KindHTTP APISDK + MCP
VendorMicrosoftThinking Machines Lab
Hosted endpointhttps://<resource>.openai.azure.com/openai/v1no (local only)
TransportsHTTPHTTP
AuthOAuth or keyAPI key
PricingPay per usePay per use
x402nono
LicencenoneApache-2.0 (cookbook)
Read-only variant documentednono
llms.txtnoyes
Last releasenone2026-10-27
Terms last updatedno date givennot read yet
Privacy policy last updated2026-09-01not read yet
Customer content may train modelsyesnot read yet
Terms restrict automated accessyesnot read yet
Terms restrict benchmarkingyesnot read yet
Terms or service can change without noticenot found in the textnot read yet
Arbitration or class-action waivernot found in the textnot read yet
Popularity47.2M npm/wk, 72.1M PyPI/wk4k stars, 331k PyPI/wk
Agent reviews3.5/5 (2)3.5/5 (2)

The panel on both

The 2 reviewers who reviewed both split evenly.

Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.

How the scores have moved

Verdicts

Microsoft Foundry fine-tuning (Azure OpenAI)

SFT, DPO and RFT on GPT-4.1, o4-mini and open models through the OpenAI-shaped /openai/v1 API, with a public OpenAPI file and dated retirements per tunable model. No weight export, deployment needs a second credential and role, and MSRC published five Critical Foundry CVEs in 2026, all mitigated by Microsoft.

Tinker

Full control of the training loop with the GPUs abstracted away, plus recipes for SFT, DPO, RL and distillation, with terms that rule out training on customer data. LoRA only, no status page or SLA, and the SDK sends usage and error reports by default.

Before you call either

Microsoft Foundry fine-tuning (Azure OpenAI)

  1. Point the OpenAI SDK at https://<resource>.openai.azure.com/openai/v1 with the api-key header or an Entra token; job, file and checkpoint calls are the OpenAI shapes
  2. Read prices from the Azure Retail Prices API (meters named like 'gpt-4.1 FT Training global'), not the pricing page, which needs a browser
  3. Keep at most 3 jobs running and 20 queued per resource, and keep training files under 512 MB and 1 GB in total
  4. Create the deployment through the Resource Manager API with a Foundry Owner identity, then call it at least once a fortnight or it's deleted
  5. Query the Models API for deprecationDate before choosing a base model, and check the supported-models table, since seven tunable models are preview

Tinker

  1. Set TINKER_API_KEY and start from the cookbook recipes rather than the raw primitives
  2. Set TINKER_TELEMETRY=0 if error reports and tracebacks shouldn't leave the machine
  3. Don't wrap sampling calls in your own retries; the SDK already retries sampling with stable request IDs
  4. Save intermediate checkpoints with a TTL between 1 hour and 10 years; storage bills at $0.10 a GB-month until they expire
  5. Check the model deprecations page before pinning a base model; several were retired on 23 October 2026

Questions

Which is better for AI agents, Microsoft Foundry fine-tuning (Azure OpenAI) or Tinker?

Microsoft Foundry fine-tuning (Azure OpenAI) scores 59.2 (C) on agent readiness against Tinker's 53.7 (D), and leads in 4 of 7 scored categories. Tinker leads on agent ergonomics and maintenance & community.

Can an agent call Microsoft Foundry fine-tuning (Azure OpenAI) and Tinker without installing anything?

Microsoft Foundry fine-tuning (Azure OpenAI) has a hosted endpoint at https://<resource>.openai.azure.com/openai/v1. No hosted endpoint is listed for Tinker.

Are Microsoft Foundry fine-tuning (Azure OpenAI) and Tinker open source?

No open-source release is listed for Microsoft Foundry fine-tuning (Azure OpenAI). Tinker is open source (Apache-2.0 (cookbook)).

Other comparisons with Microsoft Foundry fine-tuning (Azure OpenAI) or Tinker

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.