---
title: "AlphaSignal Post Training: Distillation News"
url: "https://alphasignal.ai/topic/post-training/distillation"
description: "Latest AI and machine learning news about Distillation (Post Training) on AlphaSignal."
---

# AlphaSignal Post Training: Distillation News

Latest AI and machine learning news about Distillation (Post Training) on AlphaSignal.

## Articles

- [Kokoro-82M Clones Any Voice From a 3-Second Clip for Under $20](https://alphasignal.ai/news/kokoro-82m-clones-any-voice-from-a-3-second-clip-for-under-20.md) (2026-10-07)
- [Humanizer Rewrites AI Text Locally So Detectors Flag Just 11 of 210](https://alphasignal.ai/news/humanizer-rewrites-ai-text-locally-so-detectors-flag-just-11-of-210.md) (2026-10-06)
- [Cohere's North Small Translate Beats DeepL Across 50 Languages With Downloadable Weights](https://alphasignal.ai/news/cohere-s-north-small-translate-beats-deepl-across-50-languages-with.md) (2026-10-03)
- [Amazon's ALoDLM Beats Qwen3-8B With 2.7x Faster AI Text Generation](https://alphasignal.ai/news/amazon-s-alodlm-beats-qwen3-8b-with-2-7x-faster-ai-text-generation.md) (2026-10-03)
- [Princeton's Queen Teaches a Grandmaster Chess Engine to Explain Its Moves](https://alphasignal.ai/news/princeton-s-queen-teaches-a-grandmaster-chess-engine-to-explain-its-moves.md) (2026-10-02)
- [NYU's HyperThink Cuts AI Reasoning Costs 10x Without Long Thinking Traces](https://alphasignal.ai/news/nyu-s-hyperthink-cuts-ai-reasoning-costs-10x-without-long-thinking-traces.md) (2026-10-02)
- [MIT and Tel Aviv's LSL Stops AI from Forgetting Skills After Fine-Tuning](https://alphasignal.ai/news/mit-and-tel-aviv-s-lsl-stops-ai-from-forgetting-skills-after-fine-tuning.md) (2026-10-01)
- [Kyutai's Structure-Agnostic Distillation Beats Dense Matching for Diffusion Image Models](https://alphasignal.ai/news/kyutai-s-structure-agnostic-distillation-beats-dense-matching-for-diffusion.md) (2026-09-30)
- [Simplex Diffusion Models Beat Discrete Rivals at Math With 16x Fewer Steps](https://alphasignal.ai/news/simplex-diffusion-models-beat-discrete-rivals-at-math-with-16x-fewer-steps.md) (2026-09-28)
- [Recursive Self-Distillation Doubles Qwen3-8B Math Accuracy to 65.97%](https://alphasignal.ai/news/recursive-self-distillation-doubles-qwen3-8b-math-accuracy-to-65-97.md) (2026-09-25)
- [Perplexity Turns AI Failures Into Training Data, Cutting Tool Errors 21%](https://alphasignal.ai/news/perplexity-turns-ai-failures-into-training-data-cutting-tool-errors-21.md) (2026-09-22)
- [Viggle Shrinks Qwen-Image-2.1 Generation From 40 Steps to Five](https://alphasignal.ai/news/viggle-shrinks-qwen-image-2-1-generation-from-40-steps-to-five.md) (2026-09-22)
- [Xiaomi Releases MiMo-V2.6 as a 9B Open Agent Anyone Can Run](https://alphasignal.ai/news/xiaomi-releases-mimo-v2-6-as-a-9b-open-agent-anyone-can-run.md) (2026-09-21)
- [UCL's iSDFT Keeps AI Sharp on New Tasks Without Forgetting Old Skills](https://alphasignal.ai/news/ucl-s-isdft-keeps-ai-sharp-on-new-tasks-without-forgetting-old-skills.md) (2026-09-21)
- [MIT's Boltz-2 Ranks 2x More Drug Hits With Just 300 Labels](https://alphasignal.ai/news/mit-s-boltz-2-ranks-2x-more-drug-hits-with-just-300-labels.md) (2026-09-21)
- [Bespoke Labs' Nimble Beats a 27B Model at Classification Using Just 9B](https://alphasignal.ai/news/bespoke-labs-nimble-beats-a-27b-model-at-classification-using-just-9b.md) (2026-09-20)
- [Jared Palmer's Kev-4B Skips Text Generation to Score Decisions Directly](https://alphasignal.ai/news/jared-palmer-s-kev-4b-skips-text-generation-to-score-decisions-directly.md) (2026-09-19)
- [UkisAI's Swift-Qwen3.8 Slashes AI Reasoning Tokens by 58% With Almost No Accuracy Loss](https://alphasignal.ai/news/ukisai-s-swift-qwen3-8-slashes-ai-reasoning-tokens-by-58-with-almost-no.md) (2026-09-08)
- [What Developers Can Learn From Shopify’s Self-Improving AI Pipeline](https://alphasignal.ai/news/what-developers-can-learn-from-shopifys-self-improving-ai-pipeline.md) (2026-09-07)
- [ETH Zurich's Humanoid Robot Swings Across Monkey Bars in 14 of 15 Tries](https://alphasignal.ai/news/eth-zurich-s-humanoid-robot-swings-across-monkey-bars-in-14-of-15-tries.md) (2026-08-30)
- [Alibaba's MiniMax-H3 Cuts Video Generation Steps from 32 to 8](https://alphasignal.ai/news/alibaba-s-minimax-h3-cuts-video-generation-steps-from-32-to-8.md) (2026-08-26)
- [Tencent's AuK Merges 16 Speech Tasks Into One Open-Source Model](https://alphasignal.ai/news/tencent-s-auk-merges-16-speech-tasks-into-one-open-source-model.md) (2026-08-21)
- [Community-Built LoRA Cuts Krea 2 Turbo's Image Generation Steps in Half](https://alphasignal.ai/news/community-built-lora-cuts-krea-2-turbo-s-image-generation-steps-in-half.md) (2026-08-20)
- [Tsinghua's Open-MOPD Merges Expert AI Models Into One, Recovering 83% of Lost Performance](https://alphasignal.ai/news/tsinghua-s-open-mopd-merges-expert-ai-models-into-one-recovering-83-of-lost.md) (2026-08-19)
- [Liquid AI's LFM2.5 Hits 97% Full-Precision Quality in a Tiny 4-Bit File](https://alphasignal.ai/news/liquid-ai-s-lfm2-5-hits-97-full-precision-quality-in-a-tiny-4-bit-file.md) (2026-08-19)
- [CMD Fixes the Hidden Flaw Quietly Holding Back Fast Video AI](https://alphasignal.ai/news/cmd-fixes-the-hidden-flaw-quietly-holding-back-fast-video-ai.md) (2026-08-13)
- [Salesforce AI Boosts a Weak Model's Accuracy 87% Without Touching Its Weights](https://alphasignal.ai/news/salesforce-ai-boosts-a-weak-model-s-accuracy-87-without-touching-its-weights.md) (2026-08-12)
- [Bridgewater's Fine-Tuned Model Beats GPT, Claude, and Gemini at 13.8x Lower Cost](https://alphasignal.ai/news/bridgewater-s-fine-tuned-model-beats-gpt-claude-and-gemini-at-13-8x-lower-cost.md) (2026-06-30)
- [NVIDIA's TAO 7 Lets Claude Code Train Vision AI in Plain English](https://alphasignal.ai/news/nvidia-s-tao-7-lets-claude-code-train-vision-ai-in-plain-english.md) (2026-06-30)
- [Intelligent Internet's Zenith Jumps GPT-5.5 From 5th to 1st on Hardest Coding Benchmark](https://alphasignal.ai/news/intelligent-internet-s-zenith-jumps-gpt-5-5-from-5th-to-1st-on-hardest-coding.md) (2026-06-29)
- [Automagic v2 Cuts Diffusion Model Training VRAM by 5x on Consumer GPUs](https://alphasignal.ai/news/automagic-v2-cuts-diffusion-model-training-vram-by-5x-on-consumer-gpus.md) (2026-06-23)
- [Arcee's mergekit Lets Developers Merge AI Models Without Writing Training Code](https://alphasignal.ai/news/arcee-s-mergekit-lets-developers-merge-ai-models-without-writing-training-code.md) (2026-06-17)
- [Ostris Shrinks Ideogram 4's 9B Memory Blocker Into a 109MB LoRA](https://alphasignal.ai/news/ostris-shrinks-ideogram-4-s-9b-memory-blocker-into-a-109mb-lora.md) (2026-06-16)
- [GILC Steers Biological AI Without Retraining, Beating Fine-Tuned Rivals](https://alphasignal.ai/news/gilc-steers-biological-ai-without-retraining-beating-fine-tuned-rivals.md) (2026-06-04)
- [MRepE Steers AI Code Generators Toward Readable Output Without Retraining](https://alphasignal.ai/news/mrepe-steers-ai-code-generators-toward-readable-output-without-retraining.md) (2026-06-04)
- [Zhejiang University's OPRD Trains Smarter AI 1.44x Faster Using Hidden States](https://alphasignal.ai/news/zhejiang-university-s-oprd-trains-smarter-ai-1-44x-faster-using-hidden-states.md) (2026-06-04)
- [Amazon's LadderMan Teaches Humanoid Robots to Climb at Near-Human Speed](https://alphasignal.ai/news/amazon-s-ladderman-teaches-humanoid-robots-to-climb-at-near-human-speed.md) (2026-06-04)
- [Google DeepMind's SURF Separates Audio Without Any Clean Training Data](https://alphasignal.ai/news/google-deepmind-s-surf-separates-audio-without-any-clean-training-data.md) (2026-06-03)
- [UC Santa Barbara's LoRi Closes the Silent Reasoning Gap by 12%](https://alphasignal.ai/news/uc-santa-barbara-s-lori-closes-the-silent-reasoning-gap-by-12.md) (2026-06-03)
- [NUS Discovers AI Models Already Know How They'll Be Judged](https://alphasignal.ai/news/nus-discovers-ai-models-already-know-how-they-ll-be-judged.md) (2026-06-03)
- [VISTEC's DuDi Teaches Tiny AI Models to Speak Southeast Asian Languages](https://alphasignal.ai/news/vistec-s-dudi-teaches-tiny-ai-models-to-speak-southeast-asian-languages.md) (2026-06-03)
- [Hint Tuning Cuts DeepSeek-R1 Reasoning Tokens by 66% With 1,000 Samples](https://alphasignal.ai/news/hint-tuning-cuts-deepseek-r1-reasoning-tokens-by-66-with-1-000-samples.md) (2026-06-03)
- [Google and Cornell's Sleep Paradigm Lets LLMs Learn Without Forgetting](https://alphasignal.ai/news/google-and-cornell-s-sleep-paradigm-lets-llms-learn-without-forgetting.md) (2026-06-02)
- [UCLA and Princeton's SDPG Trains Reasoning Models Faster Without Collapse](https://alphasignal.ai/news/ucla-and-princeton-s-sdpg-trains-reasoning-models-faster-without-collapse.md) (2026-06-02)
- [MIT Rewires How LLMs Fight Hallucinations Before They Ever Appear](https://alphasignal.ai/news/mit-rewires-how-llms-fight-hallucinations-before-they-ever-appear.md) (2026-06-02)
- [Fine-Tuned AI Planners Know More Than Their Outputs Reveal](https://alphasignal.ai/news/fine-tuned-ai-planners-know-more-than-their-outputs-reveal.md) (2026-06-02)
- [PGPO Borrows Fluid Physics to Stop Self-Teaching AI From Collapsing](https://alphasignal.ai/news/pgpo-borrows-fluid-physics-to-stop-self-teaching-ai-from-collapsing.md) (2026-06-02)
- [HARC Fixes the Broken Router That Ruins AI Model Merging](https://alphasignal.ai/news/harc-fixes-the-broken-router-that-ruins-ai-model-merging.md) (2026-06-02)
- [PRISM Merges CLIP, DINOv2 and SAM Into One Model, Beats All Three](https://alphasignal.ai/news/prism-merges-clip-dinov2-and-sam-into-one-model-beats-all-three.md) (2026-06-02)
- [NYU's PrimeSVT Shrinks Vision AI by 27% Without Specialized Hardware](https://alphasignal.ai/news/nyu-s-primesvt-shrinks-vision-ai-by-27-without-specialized-hardware.md) (2026-06-02)