The Circuitry
THE CIRCUITRYYour one-stop source for all tech news
  • Today
  • Feed
  • Events
  • Saved
HOMETODAYFEEDEVENTS
BOOKMARKS
RSS
© 2026 The Circuitry
About UsSourcesContactCorrectionsPrivacy
  • Today
  • Feed
  • Events
  • Saved
  • AAPL
  • MSFT
  • GOOGL
  • META
  • NVDA
  • TSLA
  • AMZN
  • INTC
  • SONY
  • TTWO
  • COIN
  • BTC
  • ETH
  • SOL
  • AAPL
  • MSFT
  • GOOGL
  • META
  • NVDA
  • TSLA
  • AMZN
  • INTC
  • SONY
  • TTWO
  • COIN
  • BTC
  • ETH
  • SOL

Feed

Ad
From The CircuitryVerified tech news, cross-checked.How we verify →
  • OpenAI cuts reasoning-extraction success to 0.2% on GPT-6.2 AstraTech
  • Scientists estimated the same work took 1.9 times longer without AITech
  • OpenAI releases a free classifier for child abuse imagesTech
  • Cairn agents found a zero-day and took control of a serverTech
  • TikTok launches an AI checker and starts a global likeness rolloutTech
  • Falcon 9 launches USSF-317 and lands the boosterTech
  • Anthropic mathematician says he believes he has a Riemann proofTech
  • Google will retire the Nest app and Assistant on home devices in 2027Tech
  • Niron Magnetics will build a $1.8 billion plant in DothanTech
  • Google's Gemini tool flags wildfire dispatches in under 30 secondsTech
  • Suspected China-linked group exploited vCenter before Nov. 19Security
Scroll for more
Verification
VERIFIEDConfidence: HIGH
Source identified
Claims cross-referenced
No discrepancies found
Sourcing
1source
OpenAI · track record
18Stories
100%Verified
1630d
All sources →
Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
Home/Tech/OpenAI introduces MedChemBench, with GPT-6.2 Astra at 84.5%
VERIFIEDBy Xavier Rivera·Friday, November 20, 2026 at 3:23 PM AST·2 min read

OpenAI introduces MedChemBench, with GPT-6.2 Astra at 84.5%

OpenAI has introduced MedChemBench, a 99-problem test of how models read molecules, explain lab results, and choose the next compound. At the xhigh setting, GPT-6.2 Astra scored 84.5%, while compound choice and property forecasts remain much weaker than structure work.

Source:OpenAI
Post
OpenAI introduces MedChemBench, with GPT-6.2 Astra at 84.5%
Ad
From The CircuitryVerified tech news, cross-checked.How we verify →
Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
TL;DRAI · 60 sec read

OpenAI has introduced MedChemBench, a 99-problem benchmark drawn from 117 papers and patents and 761 molecular structures. At the xhigh setting, GPT-6.2 Astra scored 84.5%, ahead of GPT-6 Astra at 83.3% and GPT-5.6 Sol at 58.2%. Structure reconstruction is the strong side. Compound selection and numerical property forecasts remain near half credit, and the results do not show a model can discover a new drug.

Morning Brief

Get the daily OpenAI brief

Every morning, in your inbox. What's in The Brief →

OpenAI has introduced MedChemBench, a 99-problem benchmark that scores how models read molecular structures, explain lab results, and pick the next compound to test. At the xhigh setting, GPT-6.2 Astra took 84.5% of the credit on offer.

GPT-6.2 Astra leads that shared setting. GPT-6 Astra scored 83.3%, GPT-6.1 Sol scored 82.0%, and GPT-5.6 Sol scored 58.2%. GPT-6 Astra did its best work rebuilding drawn molecules and explaining the evidence supplied in the prompt. Picking a compound and forecasting a numerical property was the weaker side for that model.
GPT-6 Astra did its best work rebuilding drawn molecules and explaining the evidence supplied in the prompt.
The set uses 117 papers and patents and 761 structures. For most items, the model sees the biological target, the stated goals, compounds already made, and a selection of lab findings. Scoring covers four jobs: correct atom links and stereochemistry, explanations that stay inside the given evidence, rankings that meet the stated requirements, and agreement with reported numbers and categories.
Comparison of how three models rebuilt one molecule, including a broken ring and a split structure
Comparison of how three models rebuilt one molecule, including a broken ring and a split structure · OpenAI
One item asks for predictions on three related PRMT5 inhibitors. That protein is studied in cancer drug work. The model gets earlier structures and results, then has to estimate cell-level blocking strength, how readily each compound crosses a cell barrier, and how fast liver enzymes break it down in an assay. Measurements withheld from the prompt later showed a split: compound 23 blocked the protein at a lower concentration, while compound 24 crossed the barrier more readily and lasted longer in the enzyme assay.
From The CircuitryThe Feed — live briefs across tech, all day.See what’s happening →
Another item requires an 18-atom ring to survive assembly. At the highest reasoning setting shared by all eleven models, GPT-6 Astra rebuilt that compound correctly in every one of ten tries. GPT-5.6 Sol and GPT-5.6 Luna missed it on every try.

OpenAI@OpenAI

View on X
The illustrated misses show a 15-atom ring from one model and two separate pieces from the other. On structure tasks as a group, Astra took 95.4% of the credit at xhigh.
The write-up leaves open the chance of earlier exposure to the source papers, and each answer was one pass with no tools, no feedback, and no new experiment.
The results do not show a model can find a new drug. OpenAI ran eleven GPT models. At maximum effort, GPT-6 Astra's full-benchmark score moved to 83.8%, while compound choice and number forecasts still collected only about half the credit.
Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
The write-up leaves open the chance of earlier exposure to the source papers, and each answer was one pass with no tools, no feedback, and no new experiment. The test measures reasoning over past evidence, not a decision inside a working lab.
Why this mattersAI · ~100 words

Tap a lens to see what this story means for you.

Morning Brief

Liked this? The Brief brings you the whole day in tech, verified, every morning.

Two minutes, free forever. What's in The Brief →

Reader-supported
DonateBuy me a coffee →Follow@thecircuitry_ →Follow@thecircuitry.to →
HELP US IMPROVE
Grand Theft Auto VILive hubUS store unlocks Thu Nov 19, 1:00 AM ASTFirst unlock: New Zealand · Wed 7:00 AM ASTLaunch hub →Image: Rockstar Games
From The Circuitry

See what’s happening right now

The Feed runs all day — short, verified briefs the moment they break.

Open the Feed →
From The Circuitry

Follow @thecircuitry_

Every story we publish, as it happens. No noise between.

Follow on X ↗On Bluesky ↗

Reader-supported

The Circuitry is a passion project I've always wanted to build, and I love the work behind it.

Running it costs real money. APIs, hosting, time. To keep improving the site and growing this into something useful for everyone, those costs have to be covered.

Any contribution is appreciated. If not, no pressure. Thanks for reading.

Buy me a coffee
openaiaimedicine
Grand Theft Auto VILive hubUS store unlocks Thu Nov 19, 1:00 AM ASTFirst unlock: New Zealand · Wed 7:00 AM ASTLaunch hub →Image: Rockstar Games
More inTech
  • Falcon 9 launches USSF-317 and lands the booster

    Tech · 8h
  • Anthropic mathematician says he believes he has a Riemann proof

    Tech · 21h
  • Google will retire the Nest app and Assistant on home devices in 2027

    Tech · 23h
SupportThe Work

The Circuitry is reader-supported. If you find the daily brief useful, you can buy me a coffee to keep it going.

Buy a coffee →
Ad
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →
From The CircuitryWhy The Circuitry

Verified tech news, cross-checked.

Every story is checked against independent sources before it posts — no rumors dressed up as fact.

How we verify →

MORE IN THIS BEAT

All Tech →
  • Tech·Nov 20

    OpenAI cuts reasoning-extraction success to 0.2% on GPT-6.2 Astra

    OpenAI has cut the average success rate of attacks that extract protected reasoning from an earlier request to 0.2% on GPT-6.2 Astra, from 54.0% on GPT-6 Astra. Current-request attacks are at 0.9%, and selected safeguard improvements are being applied back to GPT-6 Astra.

  • Tech·Nov 16

    Scientists estimated the same work took 1.9 times longer without AI

    Scientists in an OpenAI study estimated that the same work would have taken 1.9 times as long without AI. The estimates are self-reported, and 98 percent of those asked expected AI to speed up discovery.

  • Tech·Nov 16

    OpenAI releases a free classifier for child abuse images

    OpenAI has released a free image classifier that scores possible child sexual abuse material, including AI-generated images, for vetted organizations. ChatGPT also removes a flagged conversation and pauses the account for 24 hours.

  • Tech·Nov 20

    Cairn agents found a zero-day and took control of a server

    Autonomous Cairn agents found an unpatched flaw in a medical imaging app and took control of a server in a university campaign that began September 8, 2026. Amazon rates a Chinese-nexus link as moderate confidence, and the project advisory still lists no patched version.

  • Tech·Nov 20

    TikTok launches an AI checker and starts a global likeness rollout

    TikTok has launched a public AI checker for images and videos made with its own tools and has started a global opt-in likeness rollout. The checker is live without an account, while the likeness tool is limited to eligible creators and a notice does not by itself remove a post.