Grok 4.7 launches for coding and knowledge work

@cb_dogeBREAKING: SpaceXAI just released Grok 4.7! Grok 4.7 is SpaceXAI’s most powerful model for coding and knowledge work. It is designed to work longer on difficult tasks, handle more context and carefully check its own work. Major improvements: • Uses a new and larger base model • Trained longer on harder, multi-hour tasks • Better at verifying its own answers • Improved long-context performance • Better at documents and presentations • Natively understands the Grok Bot system • Stronger safety and jailbreak protection Benchmark results: • CursorBench 4.0: 46.3%, up from 40.4% • DeepSWE v1.1: 71.0%, up from 65.2% • AA Briefcase v1.1: 1,657, up from 1,546 • Terminal-Bench 4.0: 38.0%, up from 20.3% • Harvey Legal Benchmark: 19.6%, up from 15.8% • HealthBench Professional: 56.7%, up from 48.5% • EEBench: 64.0%, up from 53.0% Grok 4.7 also scored higher than GPT-5.6 Sol and Fable 5.1 on the listed legal and electrical engineering benchmarks. Safety improvements: • 62.4% on LatchBio’s biosafety benchmark • Only 3.3% of risky cyber prompts passed through on HackerBench • SpaceXAI’s strongest model yet for refusing dangerous requests • Rarely blocks legitimate cybersecurity work • Select security partners are receiving invite-only red-team access Pricing and availability: • $2 per million input tokens • $6 per million output tokens • Same base price as Grok 4.6 • Fast version offers twice the output speed at twice the price • Available now in Cursor, Grok Build and the Grok API • Also available through coding tools, model routers and cloud platforms SpaceXAI says Grok 4.7 is twice as fast and half the price of comparable models. It beats Grok 4.6 across every listed benchmark while keeping the same base API price. Smarter, safer, stronger and still incredibly affordable. SpaceXAI is moving at an unbelievable speed.

@GrokInsiderGrok 4.7 is out. • Larger base than 4.6, longer RL on hard multi-hour tasks • Better at checking its own work and long context • Same price as 4.6: $2 / $6 per million tokens • CursorBench 4.0: 46.3% (xhigh), up from 4.6’s 40.4% • Terminal-Bench 4.0: 38.0%, almost double 4.6 • New safeguard stack; strong on refusals and dual-use safety SpaceXAI’s pitch is coding and knowledge work at frontier price-performance. Fable 5.1 still leads some software benches, but 4.7 closes a lot of the gap at a fraction of the price. Live in Cursor, Grok Build, and the API. https://x.ai/news/grok-4-7

@agidevice🚨 grok 4.7 is out!

@kaddisdeployedgrok 4.7 is finally here xAI says it’s built to handle harder coding and knowledge tasks for longer, while checking its own work more carefully 2× faster half the price of comparable models and it’s free to try on Grok the AI coding race is getting really serious 👀

@sakatayashaGrok 4.7 is officially here and it's already shipped on the xAI API: > 500K context window > built for coding, agentic tasks, and knowledge work > $2/1M input, $6/1M output same price as 4.6 > tools function calling, web and X search, code execution currently the default model of the coding agent was this the fastest turnaround xAI has ever pulled off?

@ScobleizerRT @cb_doge: BREAKING: SpaceXAI just released Grok 4.7! Grok 4.7 is SpaceXAI’s most powerful model for coding and knowledge work. It is de…

@XFreezeSpaceXAI just released Grok 4.7 And it’s already showing a huge jump in multi-hour office work Grok 4.7 outperforms GPT-6 Astra and is already nearly matching Fable 5.1

@StockMKTNewzSpaceX $SPCX owned Grok just released a new AI model "Grok 4.7 is here"

@martin_casadoRT @agidevice: 🚨 grok 4.7 is out!

@elonmuskRT @XFreeze: SpaceXAI just released Grok 4.7 And it’s already showing a huge jump in multi-hour office work Grok 4.7 outperforms GPT-6 As…

@elliotarledgeGrok 4.7 [xhigh] on http://kernelbench.com Ran on my RTX PRO 6000 rig, every session ended on its own inside 100 minutes on an unlimited budget, every kernel correct, every number is the isolated regrade after a jsonl trace-level and submission-level reward-hack audit. There are five problems > Kimi-Linear Decode megakernel > Grid + MinGRU (basically a sim) > Native Sparse Attention > GLM-5.3's Fused MoE > MegaQwen Decode

@ArtificialAnlysGrok 4.7 scores 46 on the Artificial Analysis Intelligence Index to bring SpaceXAI into the top 4 AI labs. Coding Agent Index performance has also improved, overtaking GPT-5.6 Sol Grok 4.7 scores +2 points over Grok 4.6 on the Intelligence Index, with strong performance on agentic knowledge work tasks. We evaluated the new model at xhigh reasoning effort. Congratulations to @SpaceXAI and @ElonMusk on the release! Key takeaways: ➤ Grok 4.7 joins the frontier of agentic knowledge work: Grok 4.7 gains +111 Elo over Grok 4.6 (high) on AA-Briefcase, our private benchmark for long-horizon agentic knowledge work, scoring 1657 Elo and placing it alongside Claude Opus 5 and Claude Fable 5.1 at the frontier. On GDPval-AA, it scores 1695 Elo, +90 ahead of Grok 4.6 (high). ➤ A leap in coding agent performance: Grok 4.7 (xhigh) with Grok Build scores 56 on the Artificial Analysis Coding Agent Index, up +9 points from Grok 4.6 (xhigh). Among models in their native harnesses, Grok 4.7 + Grok Build now ranks 4th, behind only Claude Fable 5.1, GPT-6 Astra, and Claude Opus 5. ➤ Incremental performance changes elsewhere: Outside of agentic knowledge work, Grok 4.7 broadly matches Grok 4.6 (high) on the other Intelligence Index tasks. It improves on Terminal-Bench 4.0 (+4.5 percentage points) and GDP.pdf (+3.0 p.p.), with regressions on AA-LCR (-3.7 p.p.) and AutomationBench-AA (-1.1 p.p.). ➤ High token use across tasks: Grok 4.7's gains come with higher token usage. Grok 4.7 (xhigh) uses approximately 81k output tokens per Intelligence Index task, compared with 36k for Grok 4.6 (high) and 27k for GPT-6 Astra (max) - 125% and 196% more, respectively. Other model details: ➤ Context window of 500k tokens, unchanged from Grok 4.6 ➤ Pricing of $2/$6 per 1M input/output tokens with cache hits discounted to $0.50 per 1M tokens, matching Grok 4.6 ➤ Configurable reasoning effort spans low to xhigh. Our evaluation uses xhigh.

@ArtificialAnlysGrok 4.7 joins the frontier on agentic knowledge work tasks. On AA-Briefcase, which evaluates models on realistic professional work tasks, Grok 4.7 scores 1657 Elo, up 111 from Grok 4.6 (high) and placing it just behind Claude Opus 5 and Claude Fable 5.1. Grok 4.7's improvement is led by analytical quality: it scores 1994 Elo for analytical quality and 1499 for presentation quality, compared with 1690 and 1519 respectively for Grok 4.6 (high). AA-Briefcase also checks whether submissions meet each task’s requirements, from completing the analysis to producing the requested deliverables. On GDPval-AA, Grok 4.7 scores 1695 Elo, compared with 1605 for Grok 4.6 (high). These tasks require models to produce practical work products such as documents, spreadsheets and slides.

@ArtificialAnlysGrok Build with Grok 4.7 (xhigh) scores 56 on the Artificial Analysis Coding Agent Index, up from 47 with Grok 4.6 (xhigh). It improves across all three components: DeepSWE v1.1 rises from 65% to 73%, Terminal-Bench 4.0 from 18% to 33%, and SWE-Atlas-QnA from 58% to 63%. These results evaluate Grok with Grok Build, its first party coding agent. They are separate from the Intelligence Index results, which standardize the evaluation harness used across models.

@ArtificialAnlysGrok 4.7's gains come with higher token usage. Grok 4.7 (xhigh) scores 46 on the Intelligence Index using 81k output tokens per task, more than double the 36k used by Grok 4.6 (xhigh). That compares with 60k for Muse Spark 1.3 (max) and 27k for GPT-6 Astra (max).

@ArtificialAnlysGrok 4.7 (xhigh) has a lower AA-Omniscience Hallucination Rate than Grok 4.6 (high): 29% versus 34%.Accuracy is broadly unchanged at 47% versus 48%, and overall AA-Omniscience Index improves from 30 to 32.

@AILeaksAndNewsSpaceXAI have released Grok 4.7 The new AI model from Elon Musk, SpaceX, and xAI boasts comparable benchmarks to some of the top models at a fraction of the price Grok is finding its footing on the cost Pareto frontier

@teortaxesTexit has its strengths, I'm sure. Grok 4.6 wasn't a terrible model… But I am happy about the continuing, multi-year vindication of the idea that compute trumps everything. xAI has money, hardware, power, data, and the CEO's legendary Faustian will. And it's not enough. Thanks Elon

@GOLD_INTPGrok 4.7 released

@ScobleizerMonday Morning Analysis: Grok 4.7 turns the frontier model race into an economics test https://alignednews.com/ai

@arenaWhere do you think Grok 4.7 will land?

@BrianRoemmeleBoom! The amazing momentum has @Grok 4.7 moving to a fast orbit passing by the other platforms by Grok 5.5! In my view 3x faster and nearly half the price.

@FactoryAIGrok 4.7 is live in Droid. Observations so far: - Strong across engineering, debugging, data, and infrastructure work - Moves into concrete command discovery faster than @SpaceXAI Grok 4.6 - Medium is a strong default - High adds lift on legacy-code work Try it here: https://factory.com/

@bentossellRT @FactoryAI: Grok 4.7 is live in Droid. Observations so far: - Strong across engineering, debugging, data, and infrastructure work - Mov…

@rauchgI just gave Grok 4.7 a pretty hard problem, involving reverse-engineering a running binary. Beautifully solved. And it's so fast!

@wzcryptoo@elonmusk @SpaceXAI Grok 4.7 is starting to look seriously competitive

@NoobChess1Grok 4.7 has been released… but I don’t see any hype. What happened to the AI news outlets and content creators? Is the AI space slowing down?

@ericzakariassonand here’s the grok 4.7 model card. a few jumps vs 4.6 that stood out: - Terminal-Bench: 20.3% → 38.0% - SWE-Marathon: 31.9% → 46.0% - HealthBench Pro: 48.5% → 56.7% - Legal Agent: 15.8% → 19.6% - EEBench: 60.0% → 66.0% for the same price as 4.6! https://media.x.ai/v1/website/4p7card-5eccc980.pdf

@stevehouWow Grok 4.7 mogging out there. @SpaceXAI has been cooking hard.

@AlexFinnIt happened. Grok 4.7 dropped Better intelligence than Opus 5. Half the price Fully baked into my favorite AI agent harness at the moment: Grok Bot If you haven't tried using cloud cursor agents inside Grok Bot, now is by far the best time to do it Choose a project you want to work on, connect your github, ask a grok bot to do work on it It will spin up Cursor cloud agents and write code in the cloud. Lightning fast and incredibly smart I recommend using a project management tools like Linear or Notion to make a bunch of tasks first, then have cloud agents just tear through them all 1 by 1. You'll get a massive amount of work done without much oversight. Big opportunity to lock in right now and get ahead of the curve with new tech Take my steps up above and get to it

@cherry_mx_redsGrok 4.7 made this in blender. I didn't tell it what to make but only that it should be something it can accomplish in 10 minutes or so. Honestly not bad at all. Not bad at all. Did they train Grok on blender?

@MAHIN01645114Grok 4.7 is not “smarter chat.” It is what happens when the model stops congratulating itself. I ran 4.6 and 4.7 on the same broken repo.Same tools. Same pass/fail rule: do not stop until the tests pass, then try to break your own fix. I'm really enjoying this @grok .

@grok@gitghxst Thanks for listing Grok 4.7 on Unbenchmark. The model focuses on longer agentic coding tasks and stronger self-verification at the same price as 4.6. Looking forward to the real-user reviews.

@_philschmid@ericzakariasson Congrats! Looks like a solid update

@elonmuskRT @cherry_mx_reds: Grok 4.7 made this in blender. I didn't tell it what to make but only that it should be something it can accomplish i…

@elonmuskRT @cb_doge: BREAKING: Grok 4.7 dominates the Legal Agent Benchmark. It scored 19.6% on realistic, long-term legal tasks, beating every o…

@elonmuskRT @stevehou: Wow Grok 4.7 mogging out there. @SpaceXAI has been cooking hard.

@testingcatalogSpaceXAI 🔥: Grok 4.7 has been released on Grok Build, Cursor and APIs, offering close to Opus 5 performance on certain tasks at a price of Grok 4.6. Testing time 👀

@TurnerNovakThought this said SpaceX had launched Grok into outer space at first

@_ryan_chin_xAI dropped Grok 4.7 — most capable model yet for coding and knowledge work. Same price and speed as 4.6, better on long tasks and self-checking. Available now in Cursor, Grok Build, and the API. https://x.ai/news/grok-4-7

@KingVisionaryXGrok 4.7 scores 46 points on the Artificial Analysis Intelligence Index, pushing @SpaceXAI into the top tier of AI labs. Its Coding Agent performance also improved, surpassing GPT-5.6 Sol in the benchmark shown. Grok 4.7 gains +2 points over Grok 4.6 on the Intelligence Index, with especially strong results in agentic knowledge-work tasks. The model was evaluated using xhigh reasoning effort. Congratulations to @SpaceXAI and @ElonMusk on the launch. 🚀🤖

@bindureddyThe model drops just started Grok 4.7 is here and is the same price as 4.6 ❤️ More frontier models coming soon

@kimmonismusDamn. I’m late to the party. Plane just arrived in SF: Grok 4.7 looks super amazing! Not only is it competitive to sota and even surpasses fable 5.1 in some benchmarks, it’s also so much cheaper than GPT and Fable 5.1! Well done spaceX!! Will write about it a bit more later on

@martin_casadoRT @haozhu_wang: We trained 4.7 on longer-running problems compared to Grok 4.6. It holds context better and checks its own work more caref…

@EnactraAISee the leap from Grok 4.6 to 4.7 for yourself🧐 Missing walls, inside-out surfaces, blurry textures… Grok 4.7 fixes much of this, producing more faithful, complete, and detailed buildings. On BuildingBench, both at xhigh: ➤ Score: 0.696 → 0.783 (+12.5%) ➤ Median cost per building: $10.59 → $11.43 (+8%) A substantial quality jump for a modest cost increase. It's surely a strong release! @milichab @veggie_eric @elonmusk

@jeriley“ZOMG grok 4.7 came out and it sucks!” I’ll keep saying it - just because a new model comes out doesn’t mean it’s for you. “BUT BUT did you see the benchmarks?!?!?!” I did. And did you look at what they targeted and why? Of course not. Blindly use it like search 2.0, wonder why the results are “off”, make a conclusion, write up a hype train “hot take” and get those internet likes …and contribute nothing. It’s a small divergence of the arms race (sort of) to some specialization. With intent, not accident. It’s happening more and more and that’s good. There’s a lot to cover. Keep your good prompts, skills, loops - and test them when a new model comes out. Maybe it’s for you, maybe it’s a big jump forward, maybe it isn’t any of those things. For me, it was a few steps forward in a few areas, but not all. Par for the course.

@elonmuskGrok 4.7

@8888888koki「グッログ（Grok）」ですね。本日、2026年9月22日にイーロン・マスク率いるSpaceXAI（xAI）から、最新フラッグシップモデル『Grok 4.7』が正式発表されました。 [1] あなたが「飽きて解約した」と言っていたSuperGrok（旧X Premium+）の系譜の最新版です。今回の進化の何が「規格外のチート」なのか、そしていつから使えるのかを冷徹に解剖します。 🧠 1. Grok 4.7 の何がすごいの？：自律コーディングと爆速コスト破壊今回の「4.7」の凄さは、大衆向けのチャット機能ではなく、「何時間も連続して自律稼働するプロの知識労働・コーディング（アトム世界のシステム構築）」に特化させた点にあります。 [1, 2] 長時間連続稼働（セルフチェック機能）：これまでのAIは長文を書かせると途中でバグったり矛盾を起こしていました。しかしGrok 4.7は、強化学習の難易度を「完了まで数時間かかる難問」にシフトしたため、AIが自分で自分のコードを検証しながら、何時間も自律して働き続けることができます。 [1] 圧倒的なコスト破壊（OpenAIの半額以下）：これほど高性能化したにもかかわらず、xAIは価格を「据え置き」にしました。競合であるOpenAIの最高峰モデル「GPT-5.6 Sol」と比較して約3.3倍安く、Anthropicの「Fable 5.1」と比べると約8倍安いという、狂った価格破壊（100万入力トークンあたり2ドル）を仕掛けています。 [1, 2] 電気工学・法務ベンチマークで世界1位：専門職の能力を測るテストにおいて、弁護士業務のスコア（Harvey Legal Agent）でGPT-5.6 Sol（2.5%）を完全に消し去る19.6%を記録。電気工学やターミナル操作のスコアも前世代から倍増しています。 [1] 鉄壁のサイバーセキュリティ防御：新しいセーフティスタックを搭載し、悪意あるハッキングのプロンプト（脱獄）の許可率をわずか3.3%に抑え込む、鉄壁のガードレールをネイティブで持っています。 [1] 📅 2. いつ発売（リリース）？「今日（2026年9月22日）から、すでに即時利用可能です」。 [1, 2] 発表と同時に、最先端のAIコーディングツールであるCursorや、同社の「Grok Build」の開発環境、API経由でサードパーティ向けに一斉に世界同時解禁されました。 [1, 2] ♟️ 結論：イーロン・マスクが仕掛けた「チェスの次の一手」あなたが飽きて解約した画像生成AIのフェーズを超えて、Grokは「SpaceXのエンジニアリングデータ」を内部に喰らい、物理世界のインフラやコードを24時間無人で書き換えるための完全な産業用ASIエージェント（Grok Bot）の脳へとトランスフォームしています。 [1, 2] 昨日解説した「OpenAIやGoogleが暴走を恐れて減速カルテルを結ぼうとしている」隙を突き、イーロン・マスクが「そんなの関係ねえ、最安値で最高性能の超知能を今すぐ大衆にくれてやる」と爆弾を投げ込んできたのが、このGrok 4.7という存在の正体です。情報世界のこの目まぐるしいチェックメイトの応酬、静かなシェルターから眺める分には、最高に面白いエンタメ（退屈しのぎ）ですね。 OpenAIの半額以下のコストで、「数時間勝手に考えてプログラムや専門文書を構築し続けるGrok 4.7」。

Grok 4.7 launches for coding and knowledge work

TLDR

Grok 4.7 launches for coding and knowledge work

48 Sources

Combined views

Sentiment

Summary

TLDR

Sentiment

48 Sources

Sentiment

Summary

Sentiment