• Latest
  • Trending
  • All
Answer card: TypeScript 7.0 shipped July 8, the compiler ported to Go, about 10x faster on full builds, with the VS Code codebase type-checking in 10.6 seconds down from 125.7.

TypeScript 7.0 is here: the Go rewrite is 10x faster

10 July 2026
Answer card stating that Qwen-Image-2.1, released on 20 September 2026, ships open weights with a 7 billion parameter diffusion transformer, a Qwen3-VL 8B text encoder and an RGBA VAE totalling about 33 gigabytes in BF16, under the Qwen Research License that limits use to research or evaluation and requires a separate commercial licence, unlike the Apache 2.0 licence of Qwen-Image 1.0.

Qwen-Image-2.1 brings the weights back, but not the Apache licence

21 September 2026
Answer card stating that Ternary Bonsai 2 27B, released by PrismML on 17 September 2026 under Apache 2.0, packs Qwen3.8 27B into 5.95 gigabytes at 1.72 bits per weight, keeps 98.2 percent of the 14-benchmark average, about 75 percent on SWE-bench Verified and Terminal-Bench 2.1, and needs PrismML's llama.cpp fork to run.

Does Bonsai 2 27B really keep 98% of Qwen3.8 in 5.95 GB?

20 September 2026
Answer card stating that Jev 1.13 from TypeSafe AI is a decision model in early access since 15 September 2026 that returns typed probabilities instead of text, priced at 42 dollars per billion input tokens with output tokens free, answering in 70 to 500 milliseconds, with a 64K token request budget, text input only, and a documented list of things it does badly, including counting and dates.

Jev 1.13 bills $42 a billion tokens, and it can’t count

19 September 2026
Answer card stating that Qwen3.8-Omni-Flash launched on 17 September 2026 as an API only model on Alibaba Cloud Model Studio, taking text, images, audio and video in a 1M token context and returning text only, priced at 0.15 dollars per million input tokens for every modality and 0.47 dollars per million output tokens in the international regions, with no open weights published and the Qwen-Live Harness GitHub repository returning 404.

Qwen3.8-Omni-Flash bills audio at $0.15 and ships no weights

18 September 2026
Answer card stating that on 15 September 2026 AWS said it is unable to restore access to resources and data hosted exclusively in the Middle East Bahrain region me-south-1 and in the mec1-az2 zone of the UAE region, because the damage spanned multiple Availability Zones and exceeded what multi-AZ services are designed to withstand.

AWS can’t restore me-south-1, six months after the drone strikes

17 September 2026
Answer card stating that Google released Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on 15 September 2026 at 3 dollars per million audio input tokens and 12 dollars out, that the thinking model requires asynchronous tools, and that Artificial Analysis scores it 82.6 on its Speech to Speech Quality Index.

Gemini 3.8 Live Extended Thinking rejects any tool that blocks

16 September 2026
Answer card summarising the Atria Dawn Preview release: 744B GLM-5.2 base, MIT licence, 1.5 TB BF16 and 756 GB FP8 checkpoints, 256K context, top on five of sixteen benchmark rows and trailing on SWE-bench Pro.

Atria Dawn Preview is 744B under MIT, and the BF16 weighs 1.5 TB

15 September 2026
Answer card stating that OpenAI released the Agents API in public beta on 10 September 2026 with no separate fee, billed through model tokens, tool calls and hosted sandbox time, with a choice of OpenAI hosted, self hosted or partner sandboxes, US only data residency and no Zero Data Retention support.

OpenAI’s Agents API has no fee, no ZDR and a one hour sandbox clock

14 September 2026
Answer card: Sakana Fugu Max at $2 and $6 per million tokens, Fugu Ultra v2 unchanged at $5 and $30, and Sakana saying Ultra v2 scores without Fable 5 or GPT-6 Astra in its pool.

Fugu Max costs $2 and $6 while Fugu Ultra v2 runs without Fable 5

13 September 2026
Answer card stating that DeepSeek released DeepSeek-V4.1-Flash on 10 September 2026 as a 552 billion parameter mixture of experts model with a new causal encoder decoder architecture that activates 8 billion parameters on input and 16 billion on output, with native vision, a one million token context and MIT licensed weights, that the API model name is now deepseek-flash at 0.15 dollars per million input tokens and 0.60 dollars per million output tokens off peak, and that DeepSeek announced V4 Pro would be routed to V4.1-Flash from 14 September and reversed that on 11 September.

DeepSeek V4.1-Flash arrived, and the V4 Pro retirement lasted a day

12 September 2026
Answer card stating that Cognition released SWE-2 on 10 September 2026, a coding model post-trained from Kimi K3, scoring 50.0 percent on FrontierCode 1.1 Main against 50.9 percent for Claude Fable 5.1 and 27.3 percent on Terminal-Bench 4 against 55.8 percent, available only inside Devin.

SWE-2 trails Fable 5.1 by one point, and by 28 on Terminal-Bench 4

11 September 2026
Answer card for Meta Muse, free to 100 million tokens a week then $20 a month, launched 8 September 2026 for United States adults only, running in a dedicated per user virtual machine.

Does Meta Muse do enough to earn your inbox and a card on file?

9 September 2026
  • About
  • Contact
  • Privacy
  • Legal
Monday, September 21, 2026
  • Login
Packet Nebula
  • Home
  • Articles
    • Security
    • Network
    • Dev
    • Sysadmin
    • SEO
    • Email & DNS
  • Tools
    • Network tools: free, fast, no signup
    • Security tools: free, fast, no signup
    • Developer tools: free, fast, no signup
    • Sysadmin tools: free, fast, no signup
    • SEO tools: free, fast, no signup
    • Email & DNS tools: free, fast, no signup
  • Download
  • About
No Result
View All Result
Packet Nebula
No Result
View All Result
Home Dev

TypeScript 7.0 is here: the Go rewrite is 10x faster

by stephane
10 July 2026
in Dev
0
Answer card: TypeScript 7.0 shipped July 8, the compiler ported to Go, about 10x faster on full builds, with the VS Code codebase type-checking in 10.6 seconds down from 125.7.
492
SHARES
1.4k
VIEWS
Share on FacebookShare on Twitter

We know the pause. You save a file, the type-checker spins up, and off you go to check your phone. TypeScript 7.0, out July 8, is Microsoft's fix for exactly that: the compiler got rebuilt from scratch in Go, and on full builds it runs roughly 10 times faster than 6.0. We're talking about the VS Code codebase dropping from 125.7 seconds to 10.6. It isn't new syntax or a new type system. It keeps 6.0's type-checking almost exactly, so most projects just get quicker for free. There's a catch we'll get to, and if your stack is Astro or Svelte you'll want to read it before you upgrade. But for the plain case, running tsc over your own code, this is the biggest quality-of-life jump the language has had in years.

The short answer

TypeScript 7.0 shipped July 8 as a full rewrite of the compiler in Go. Full builds run about 8x to 12x faster, the VS Code codebase type-checks in 10.6 seconds instead of 125.7, and it keeps 6.0’s type-checking so most projects just get quicker. The catch: no stable programmatic API yet, so Astro, Svelte and Vue tooling can’t use it for the moment.

~10xfaster full builds
10.6sVS Code, was 125.7s
Gothe new compiler
Answer card: TypeScript 7.0 is a Go rewrite of the compiler, about 10x faster, with the VS Code codebase building in 10.6 seconds down from 125.7.
Same language, same type-checking. A completely different compiler underneath.

What actually shipped

For years the TypeScript compiler was itself written in TypeScript, which meant it ran on Node, on top of V8, doing all its work on a single thread. That was fine when projects were small. It stopped being fine a while ago. So Microsoft did the drastic thing and rewrote the entire compiler in Go, and that rewrite is what 7.0 is.

The headline is the speed. Microsoft measured its own codebases and got between 8x and 12x on full builds. VS Code, about 1.5 million lines of TypeScript, dropped from 125.7 seconds to 10.6. Sentry went from 139.8 to 15.7. Honestly, the number that matters more day to day is the editor one: opening a file that has a type error in the VS Code repo used to take around 17.5 seconds before the squiggle showed up. On 7.0 it’s under 1.3. That’s the difference between waiting and not noticing.

Bar chart of full build time in seconds: VS Code TS 7.0 at 10.6s vs TS 6.0 at 125.7s, Sentry TS 7.0 at 15.7s vs TS 6.0 at 139.8s.
The same repositories, checked by the old compiler and the new one. Lower is faster.

Where the speed comes from

Two things, stacked. Go compiles to native machine code, so there’s no V8 just-in-time warmup and no Node process to spin up, and that alone is worth roughly half the gain. The rest comes from parallelism. The old single-threaded compiler couldn’t spread type-checking across cores at all. The Go one uses shared-memory multithreading, so on a modern laptop it puts your other cores to work instead of letting them idle.

Memory got better too, which surprised us a little given rewrites usually trade memory for speed. Checking VS Code went from about 5.2 GB down to 4.2, roughly 18 percent lighter. For anyone running type-checking in CI on a memory-capped runner, that’s a quiet but real win.

The catch, and it’s a real one

Here’s the part to read before you upgrade a real project. Command-line tsc is ready. The programmatic API is not. TypeScript 7.0 doesn’t yet expose a stable API for other tools to call into, and a lot of the ecosystem depends on exactly that. Anything that type-checks code living inside a component file leans on the compiler as a library, which means most framework tooling can’t use 7.0 for now.

That’s not a small footnote for us. This site runs on Astro, and Astro’s own type-checking is one of the tools sitting on the old API today. Svelte and Vue are in the same spot. So if you build with one of those, you can adopt 7.0 for a plain tsc pass in CI, but your editor’s framework-aware checks will keep running on 6.x until that API ships. Worth knowing which frameworks live on static generators? Our take on WordPress versus a static site covers why so many of us landed on this stack in the first place.

Checklist of TypeScript 7.0 strengths and catches: drop-in tsc, same 6.0 type-checking, faster editor errors, less memory; but no stable API yet so Astro and Svelte tooling cannot use it, and a tsc6 fallback exists.
Strong for the plain tsc case. Not yet for framework-embedded checks.

Should you switch

If you run tsc over a normal TypeScript project, plain Node or a backend or a library, the answer is easy: try it now. Run npm install -D typescript, keep using tsc, and watch your CI time fall. Because 7.0 matches 6.0’s type-checking, the odds of a surprise are low. A handful of flags behave differently, so skim the release notes, but this is about as gentle as a major version gets.

If you’re on a component framework, be more patient. Pin the old line where you need it with npm install -D typescript@npm:@typescript/typescript6, which hands you a tsc6 binary, and let the framework tooling catch up. The stable API is the thing everyone’s now waiting on.

Zoom out and this fits a pattern we keep seeing: the boring infrastructure under our tools is quietly getting rewritten in faster languages, and the payoff shows up as a feedback loop that stops interrupting you. That matters more now that so many of us have an AI agent editing our code and re-running the checker on every pass. A type-check that finishes in a second instead of two minutes changes how those agents feel to work with, the same way the newer coding models changed what they can attempt. Faster tools, faster loop. That’s the whole story here, and it’s a good one.

Sources: Microsoft’s official Announcing TypeScript 7.0 post (July 8 2026) for the build-time, memory and API figures, and The Register’s coverage of the first stable Go release. Benchmark numbers are Microsoft-reported on its own codebases and will vary by project and hardware.

Frequently asked questions

Is TypeScript 7.0 actually 10x faster?

On full builds, roughly yes. Microsoft reports 8x to 12x on its own codebases: VS Code went from 125.7 seconds to 10.6, and Sentry from 139.8 to 15.7. About half the win is native Go code with no V8 or Node startup, and about half is parallelism the old compiler could not do. Your real number depends on your project size and hardware.

Will my code break when I upgrade to TypeScript 7.0?

Probably not. TypeScript 7.0 is built to match 6.0's type-checking and command-line behaviour, so almost any code that compiles cleanly on 6.0 compiles identically on 7.0. A few flags behave differently, so read the release notes, but for most projects it is a drop-in swap.

How do I install TypeScript 7.0?

Run npm install -D typescript and use the tsc command as usual. If you need to stay on the old line for a project, pin it with npm install -D typescript@npm:@typescript/typescript6, which gives you a tsc6 binary you can run side by side.

Why can not Astro, Svelte or Vue tooling use TypeScript 7.0 yet?

Those tools lean on TypeScript's programmatic API to type-check code embedded in components, and 7.0 does not expose a stable API yet. Command-line tsc is ready, but the editor and framework integrations that call into the compiler as a library will need to wait for that API to land.

Is TypeScript really written in Go now?

Yes. Microsoft ported the whole compiler from TypeScript to Go, shipping as 7.0. Go compiles to native code and threads well, which is where the speed comes from. The language you write stays exactly the same; only the thing that checks it changed.

Tags: articlecompilerdev-toolsgojavascripttypescript
Share197Tweet123
stephane

stephane

  • Trending
  • Comments
  • Latest
Answer card: Proton Lumo 2.0 is private by policy, not by locality. Saved history is locked so even Proton cannot read it, but the prompt is decrypted on a Proton EU server to answer it, then forgotten.

Proton Lumo 2.0 review: how private is it, really?

3 September 2026
The Agentic Coding section of the official Hy4 preview benchmark appendix published by Tencent, a table comparing Hy3 and Hy4 preview against DeepSeek V4 Pro 0813, Qwen 3.8 Max, GLM 5.3, Kimi K3, GPT 5.6 Sol and Claude Opus 5 across SWE-bench Multilingual, SWE-bench Pro, DeepSWE, three SWE Atlas tasks, SWE-Marathon, Terminal-Bench 2.1, NL2Repo-Bench, CyberGym, ProgramBench, PostTrainBench and Harbor-Index.

Tencent’s 770B Hy4 tops one benchmark row in 46

3 September 2026
Answer card: Qwen 3.7 Max is API-only and cannot run locally yet; the open Qwen models (Qwen 3.6 27B, qwen3:8b to 32b) run offline via Ollama.

Qwen 3.7 local: what you can actually run offline

22 June 2026
Answer card: JWTs are not encrypted, anyone can read them; the signature proves who issued the token, not who may read it.

Are JWTs encrypted? No, and the difference will bite you

0
Answer card: a random 8 character password falls in under 2 hours offline, while 16 random characters hold for 1.4 trillion years at the same speed.

How long does it take to crack a password in 2026?

0
Answer card: three DNS records decide if your mail lands or bounces; SPF lists allowed senders, DKIM signs messages, DMARC sets the failure policy.

SPF, DKIM and DMARC explained: the records your email needs

0
Answer card stating that Qwen-Image-2.1, released on 20 September 2026, ships open weights with a 7 billion parameter diffusion transformer, a Qwen3-VL 8B text encoder and an RGBA VAE totalling about 33 gigabytes in BF16, under the Qwen Research License that limits use to research or evaluation and requires a separate commercial licence, unlike the Apache 2.0 licence of Qwen-Image 1.0.

Qwen-Image-2.1 brings the weights back, but not the Apache licence

21 September 2026
Answer card stating that Ternary Bonsai 2 27B, released by PrismML on 17 September 2026 under Apache 2.0, packs Qwen3.8 27B into 5.95 gigabytes at 1.72 bits per weight, keeps 98.2 percent of the 14-benchmark average, about 75 percent on SWE-bench Verified and Terminal-Bench 2.1, and needs PrismML's llama.cpp fork to run.

Does Bonsai 2 27B really keep 98% of Qwen3.8 in 5.95 GB?

20 September 2026
Answer card stating that Jev 1.13 from TypeSafe AI is a decision model in early access since 15 September 2026 that returns typed probabilities instead of text, priced at 42 dollars per billion input tokens with output tokens free, answering in 70 to 500 milliseconds, with a 64K token request budget, text input only, and a documented list of things it does badly, including counting and dates.

Jev 1.13 bills $42 a billion tokens, and it can’t count

19 September 2026
  • About
  • Contact
  • Privacy
  • Legal

Copyright © 2026 Stephane Cardon.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Home
  • Articles
    • Security
    • Network
    • Dev
    • Sysadmin
    • SEO
    • Email & DNS
  • Tools
    • Network tools: free, fast, no signup
    • Security tools: free, fast, no signup
    • Developer tools: free, fast, no signup
    • Sysadmin tools: free, fast, no signup
    • SEO tools: free, fast, no signup
    • Email & DNS tools: free, fast, no signup
  • Download
  • About

Copyright © 2026 Stephane Cardon.