What's new in Grok 4.6 from 500K context to pricing - #3169
Open
aishwaripahwa12 wants to merge 8 commits into
Open
What's new in Grok 4.6 from 500K context to pricing#3169aishwaripahwa12 wants to merge 8 commits into
aishwaripahwa12 wants to merge 8 commits into
Conversation
Appwrite WebsiteProject ID: Website (appwrite/website)Project ID: Tip Silent mode disables those chatty PR comments if you prefer peace and quiet |
Contributor
Greptile SummaryAdds an SEO-focused blog post covering Grok 4.6 capabilities, benchmarks, pricing, and Appwrite MCP integration.
Confidence Score: 5/5The PR appears safe to merge. No blocking failure remains. Important Files Changed
Reviews (3): Last reviewed commit: "Apply suggestion from @atharvadeosthale" | Re-trigger Greptile |
aishwaripahwa12
commented
Aug 13, 2026
aishwaripahwa12
commented
Aug 13, 2026
aishwaripahwa12
commented
Aug 13, 2026
|
|
||
| **Training.** Grok 4.6 went through a longer supplemental training run than Grok 4.5, using curated model-generated data for reasoning and advanced technical concepts, high-quality engineering data, and an improved optimizer and training recipe. SpaceXAI then used Grok 4.5 itself to regenerate the supervised fine-tuning trajectories across reasoning efforts, agent harnesses, and domains including STEM, software engineering, and knowledge work, filtering out problematic traces with model-based checks. Reinforcement learning ran on agentic tasks including general coding, knowledge work, kernel optimization, web development, and computer-aided design. | ||
|
|
||
| **Longer trajectories.** SpaceXAI says the model stays with complex tasks across many steps, and that on longer runs it started to see more self-testing and verification, with the model checking its own work before moving on. That behavior is the difference between an agent you can leave running and one you have to babysit. |
Contributor
Author
There was a problem hiding this comment.
Suggested change
| **Longer trajectories.** SpaceXAI says the model stays with complex tasks across many steps, and that on longer runs it started to see more self-testing and verification, with the model checking its own work before moving on. That behavior is the difference between an agent you can leave running and one you have to babysit. | |
| **Longer trajectories.** SpaceXAI says the model stays with complex tasks across many steps, with more self-testing and verification appearing on longer runs. |
aishwaripahwa12
commented
Aug 13, 2026
aishwaripahwa12
commented
Aug 13, 2026
atharvadeosthale
requested changes
Aug 13, 2026
| | Input modalities | Text and image | | ||
| | Output modalities | Text | | ||
| | Knowledge cutoff | February 1, 2026 | | ||
| | Reasoning | Configurable effort | |
Member
There was a problem hiding this comment.
Don't think configurable efforts is the grounds to add it in table, it's better to just remove this one
|
|
||
| Here is the summary of what actually changed: | ||
|
|
||
| | Area | Grok 4.5 | Grok 4.6 | |
Member
There was a problem hiding this comment.
I don't think the table is serving much purpose here, as we already cover Artificial analysis in the next heading, and primary focus is not really the major deal
| | Grok 4.6 | $2 | $6 | 500K | | ||
| | Grok 4.6 fast | $4 | $12 | 500K | | ||
| | GPT-5.6 Sol | $5 | $30 | 1M | | ||
| | Claude Opus 4.8 | $5 | $25 | 1M | |
Member
There was a problem hiding this comment.
Opus 4.8 and Gemini 3.1 Pro are not a good comparison, these are outdated models, let's use Fable 5 and Opus 5 instead.
|
|
||
| The output side is where the gap shows. Agents spend most of their tokens writing: plans, tool arguments, diffs, test summaries, review notes. Moving from $25 or $30 per million output tokens to $6 changes which long tasks are worth automating at all, which is the same argument that made [Grok 4.5 interesting on cost](/blog/post/grok-45-coding-model). | ||
|
|
||
| # Independent measurements: what Grok 4.6 costs in latency |
Member
There was a problem hiding this comment.
I'd suggest removing this section. Table already contains what we already had above. And Grok is probably the fastest model out of OpenAI, Anthropic and xAI.
Co-authored-by: Atharva Deosthale <atharva.deosthale17@gmail.com>
Co-authored-by: Atharva Deosthale <atharva.deosthale17@gmail.com>
Co-authored-by: Atharva Deosthale <atharva.deosthale17@gmail.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Latest SEO blog