• bitcoinBitcoin(BTC)$60,776.001.35%
  • ethereumEthereum(ETH)$1,713.812.06%
  • tetherTether(USDT)$1.000.04%
  • binancecoinBNB(BNB)$556.871.77%
  • usd-coinUSDC(USDC)$1.000.03%
  • rippleXRP(XRP)$0.93-0.68%
  • solanaSolana(SOL)$70.37-0.27%
  • tronTRON(TRX)$0.304826-0.04%
  • Figure HelocFigure Heloc(FIGR_HELOC)$1.062.67%
  • HyperliquidHyperliquid(HYPE)$63.184.18%
  • zcashZcash(ZEC)$797.131.47%
  • moneroMonero(XMR)$489.401.95%
  • dogecoinDogecoin(DOGE)$0.0551600.07%
  • leo-tokenLEO Token(LEO)$9.170.34%
  • whitebitWhiteBIT Coin(WBT)$58.371.38%
  • chainlinkChainlink(LINK)$8.822.75%
  • USDSUSDS(USDS)$1.00-0.02%
  • cardanoCardano(ADA)$0.144546-0.29%
  • Ethena USDeEthena USDe(USDE)$1.000.04%
  • stellarStellar(XLM)$0.127621-1.00%
  • daiDai(DAI)$1.000.00%
  • bitcoin-cashBitcoin Cash(BCH)$197.75-0.17%
  • nearNEAR Protocol(NEAR)$3.005.90%
  • litecoinLitecoin(LTC)$40.03-0.09%
  • avalanche-2Avalanche(AVAX)$6.680.89%
  • CantonCanton(CC)$0.072178-5.89%
  • tether-goldTether Gold(XAUT)$3,552.420.23%
  • Ripple USDRipple USD(RLUSD)$1.000.00%
  • uniswapUniswap(UNI)$4.72-0.21%
  • USD1USD1(USD1)$1.000.05%
  • the-open-networkGram (prev. Toncoin)(GRAM)$0.981.54%
  • suiSui(SUI)$0.661.56%
  • Global DollarGlobal Dollar(USDG)$1.000.01%
  • Ondo US Dollar YieldOndo US Dollar Yield(USDY)$1.15-0.01%
  • paypal-usdPayPal USD(PYUSD)$1.00-0.03%
  • RainRain(RAIN)$0.003222-20.16%
  • hedera-hashgraphHedera(HBAR)$0.051443-0.10%
  • BittensorBittensor(TAO)$190.002.16%
  • shiba-inuShiba Inu(SHIB)$0.000004-0.61%
  • crypto-com-chainCronos(CRO)$0.040604-1.43%
  • quant-networkQuant(QNT)$142.700.03%
  • Spiko Amundi Overnight Swap Fund (EUR)Spiko Amundi Overnight Swap Fund (EUR)(EURSAFO)$1.120.00%
  • Circle USYCCircle USYC(USYC)$1.140.00%
  • okbOKB(OKB)$82.674.62%
  • MemeCoreMemeCore(M)$0.75-2.22%
  • Falcon USDFalcon USD(USDF)$1.000.04%
  • United StablesUnited Stables(U)$1.000.10%
  • pax-goldPAX Gold(PAXG)$3,553.700.15%
  • BlackRock USD Institutional Digital Liquidity FundBlackRock USD Institutional Digital Liquidity Fund(BUIDL)$1.000.00%
  • aaveAave(AAVE)$98.69-0.27%
TradePoint.io
  • Main
  • AI & Technology
  • Stock Charts
  • Market & News
  • Business
  • Finance Tips
  • Trade Tube
  • Blog
  • Shop
No Result
View All Result
TradePoint.io
No Result
View All Result

Thinking Machines Launches Tinker: A Low-Level Training API that Abstracts Distributed LLM Fine-Tuning without Hiding the Knobs

October 3, 2025
in AI & Technology
Reading Time: 5 mins read
A A
Thinking Machines Launches Tinker: A Low-Level Training API that Abstracts Distributed LLM Fine-Tuning without Hiding the Knobs
ShareShareShareShareShare

Thinking Machines has released Tinker, a Python API that lets researchers and engineers write training loops locally while the platform executes them on managed distributed GPU clusters. The pitch is narrow and technical: keep full control of data, objectives, and optimization steps; hand off scheduling, fault tolerance, and multi-node orchestration. The service is in private beta with a waitlist and starts free, moving to usage-based pricing “in the coming weeks.”

Alright, but tell me what it is?

Tinker exposes low-level primitives—not high-level “train()” wrappers. Core calls include forward_backward, optim_step, save_state, and sample, giving users direct control over gradient computation, optimizer stepping, checkpointing, and evaluation/inference inside custom loops. A typical workflow: instantiate a LoRA training client against a base model (e.g., Llama-3.2-1B), iterate forward_backward/optim_step, persist state, then obtain a sampling client to evaluate or export weights.

YOU MAY ALSO LIKE

The Winners Will Be Measured in Hours Without Human Help – Unite.AI

Which Is Better for Managing Work? – Unite.AI

https://thinkingmachines.ai/tinker/

Key Features

  • Open-weights model coverage. Fine-tune families such as Llama and Qwen, including large mixture-of-experts variants (e.g., Qwen3-235B-A22B).
  • LoRA-based post-training. Tinker implements Low-Rank Adaptation (LoRA) rather than full fine-tuning; their technical note (“LoRA Without Regret”) argues LoRA can match full FT for many practical workloads—especially RL—under the right setup.
  • Portable artifacts. Download trained adapter weights for use outside Tinker (e.g., with your preferred inference stack/provider).

What runs on it?

The Thinking Machines team positions Tinker as a managed post-training platform for open-weights models from small LLMs up to large mixture-of-experts systems, a good example would be Qwen-235B-A22B as a supported model. Switching models is intentionally minimal—change a string identifier and rerun. Under the hood, runs are scheduled on Thinking Machines’ internal clusters; the LoRA approach enables shared compute pools and lower utilization overhead.

https://thinkingmachines.ai/tinker/

Tinker Cookbook: Reference Training Loops and Post-Training Recipes

To reduce boilerplate while keeping the core API lean, the team published the Tinker Cookbook (Apache-2.0). It contains ready-to-use reference loops for supervised learning and reinforcement learning, plus worked examples for RLHF (three-stage SFT → reward modeling → policy RL), math-reasoning rewards, tool-use / retrieval-augmented tasks, prompt distillation, and multi-agent setups. The repo also ships utilities for LoRA hyperparameter calculation and integrations for evaluation (e.g., InspectAI).

Who’s already using it?

Early users include groups at Princeton (Gödel prover team), Stanford (Rotskoff Chemistry), UC Berkeley (SkyRL, async off-policy multi-agent/tool-use RL), and Redwood Research (RL on Qwen3-32B for control tasks).

Tinker is private beta as of now with waitlist sign-up. The service is free to start, with usage-based pricing planned shortly; organizations are asked to contact the team directly for onboarding.

My thoughts/ comments

I like that Tinker exposes low-level primitives (forward_backward, optim_step, save_state, sample) instead of a monolithic train()—it keeps objective design, reward shaping, and evaluation in my control while offloading multi-node orchestration to their managed clusters. The LoRA-first posture is pragmatic for cost and turnaround, and their own analysis argues LoRA can match full fine-tuning when configured correctly, but I’d still want transparent logs, deterministic seeds, and per-step telemetry to verify reproducibility and drift. The Cookbook’s RLHF and SL reference loops are useful starting points, yet I’ll judge the platform on throughput stability, checkpoint portability, and guardrails for data governance (PII handling, audit trails) during real workloads.

Overall I prefer Tinker’s open, flexible API: it lets me customize open-weight LLMs via explicit training-loop primitives while the service handles distributed execution. Compared with closed systems, this preserves algorithmic control (losses, RLHF workflows, data handling) and lowers the barrier for new practitioners to experiment and iterate.


Check out the Technical details and Sign up for our waitlist here. If you’re a university or organization looking for wide scale access, contact [email protected]. 

Feel free to check out our GitHub Page for Tutorials, Codes and Notebooks. Also, feel free to follow us on Twitter and don’t forget to join our 100k+ ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.

The post Thinking Machines Launches Tinker: A Low-Level Training API that Abstracts Distributed LLM Fine-Tuning without Hiding the Knobs appeared first on MarkTechPost.

Credit: Source link

ShareTweetSendSharePin

Related Posts

The Winners Will Be Measured in Hours Without Human Help – Unite.AI
AI & Technology

The Winners Will Be Measured in Hours Without Human Help – Unite.AI

November 22, 2026
Which Is Better for Managing Work? – Unite.AI
AI & Technology

Which Is Better for Managing Work? – Unite.AI

November 22, 2026
What Does The Blue Snowflake Icon Mean For Your Tesla Battery?
AI & Technology

What Does The Blue Snowflake Icon Mean For Your Tesla Battery?

November 22, 2026
Google AI Previews Tool Retrieval in the Gemini API for Large Tool Libraries
AI & Technology

Google AI Previews Tool Retrieval in the Gemini API for Large Tool Libraries

November 22, 2026
Next Post
LIVE: Trump and Hegseth deliver remarks to gathering of generals | NBC News

LIVE: Trump and Hegseth deliver remarks to gathering of generals | NBC News

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Search

No Result
View All Result
Nvidia Names Colette Kress as Interim CEO

Nvidia Names Colette Kress as Interim CEO

November 19, 2026
Voyager 1 hit one light day but isn't sending home any data right now – Mashable

Voyager 1 hit one light day but isn't sending home any data right now – Mashable

November 22, 2026
Treasury Secretary Scott Bessent says Nvidia should look outside the company for Jensen Huang’s successor

Treasury Secretary Scott Bessent says Nvidia should look outside the company for Jensen Huang’s successor

November 15, 2026

About

Learn more

Our Services

Legal

Privacy Policy

Terms of Use

Bloggers

Learn more

Article Links

Contact

Advertise

Ask us anything

©2020- TradePoint.io - All rights reserved!

Tradepoint.io, being just a publishing and technology platform, is not a registered broker-dealer or investment adviser. So we do not provide investment advice. Rather, brokerage services are provided to clients of Tradepoint.io by independent SEC-registered broker-dealers and members of FINRA/SIPC. Every form of investing carries some risk and past performance is not a guarantee of future results. “Tradepoint.io“, “Instant Investing” and “My Trading Tools” are registered trademarks of Apperbuild, LLC.

This website is operated by Apperbuild, LLC. We have no link to any brokerage firm and we do not provide investment advice. Every information and resource we provide is solely for the education of our readers. © 2020 Apperbuild, LLC. All rights reserved.

No Result
View All Result
  • Main
  • AI & Technology
  • Stock Charts
  • Market & News
  • Business
  • Finance Tips
  • Trade Tube
  • Blog
  • Shop

© 2023 - TradePoint.io - All Rights Reserved!