Grok 4.6 vs Grok 4.7
Different efforts: medium · low · Compare all at high
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Grok 4.663.2
Grok 4.756.7
GDPval-AA (work tasks) 2.1 GDPval-AA (work tasks) 2.1: Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work
Grok 4.655.6
Grok 4.754.4
AA Analyst Agent AA Analyst Agent: Data-analyst tasks: the model explores data with tools and reports findings
Grok 4.6 · high41.3
Grok 4.7 · high55
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Grok 4.613.1
Grok 4.716.2
Terminal-Bench Science 0.1 Terminal-Bench Science 0.1: Agents complete expert-authored research workflows across five scientific domains in a terminal
Grok 4.6 · high6.2
Grok 4.74.8
AA-Briefcase 1.1 · Elo AA-Briefcase 1.1 · Elo: Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating
Grok 4.61,483.2
Grok 4.71,498.6
AA-LCR (long context) 1.1 AA-LCR (long context) 1.1: Questions that need reading and connecting facts across very long documents (around 100K tokens)
Grok 4.681
Grok 4.780
DeepSWE 1.1 DeepSWE 1.1: Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition
Grok 4.667.5
Grok 4.7 · xhigh33.3
Furniture Assembly Furniture Assembly: Models inspect furniture assembly photos, identify incorrect builds and locate the mistaken step
Grok 4.6 · xhigh40
Grok 4.7 · xhigh20.8
τ³-Banking · AA τ³-Banking · AA: Agents retrieve banking knowledge and execute tools to resolve customer-support workflows
Grok 4.644.3
Grok 4.7—
ITBench SRE ITBench SRE: IT incident tasks: diagnose and fix broken systems like a site-reliability engineer
Grok 4.6—
Grok 4.7 · xhigh42.1
EnterpriseOps-Gym-AA EnterpriseOps-Gym-AA: Agents complete stateful enterprise workflows in eight domains; final database state determines success
Grok 4.6 · high48.3
Grok 4.7—
FrontierCode 1.1 FrontierCode 1.1: Agents implement changes in real repositories, graded against maintainer-defined requirements
Grok 4.6—
Grok 4.739.6
Harvey LAB-AA 1.1 Harvey LAB-AA 1.1: Legal agents work from case documents; every rubric criterion must pass and material hallucinations receive no credit
Grok 4.6—
Grok 4.7 · xhigh9.4
EBR Bench EBR Bench: Agents learn to play Earthborne Rangers across repeated playthroughs
Grok 4.6 · xhigh30.5
Grok 4.7—
| Benchmark | Grok 4.6 | Grok 4.7 |
|---|---|---|
| AutomationBench-AA | 63.2 | 56.7 |
| GDPval-AA (work tasks) 2.1 | 55.6 | 54.4 |
| AA Analyst Agent | 41.3 · high | 55 · high |
| Terminal-Bench 4.0 | 13.1 | 16.2 |
| Terminal-Bench Science 0.1 | 6.2 · high | 4.8 |
| AA-Briefcase 1.1 · Elo | 1,483.2 | 1,498.6 |
| AA-LCR (long context) 1.1 | 81 | 80 |
| DeepSWE 1.1 | 67.5 | 33.3 · xhigh |
| Furniture Assembly | 40 · xhigh | 20.8 · xhigh |
| τ³-Banking · AA | 44.3 | — |
| ITBench SRE | — | 42.1 · xhigh |
| EnterpriseOps-Gym-AA | 48.3 · high | — |
| FrontierCode 1.1 | — | 39.6 |
| Harvey LAB-AA 1.1 | — | 9.4 · xhigh |
| EBR Bench | 30.5 · xhigh | — |
Specs
| Spec | Grok 4.6 | Grok 4.7 |
|---|---|---|
| Input, per 1M | $2 | $2 |
| Output, per 1M | $6 | $6 |
| Cost per task? | $1.14 | $1.25 |
| Speed, tok/s | 68 | 80 |
| Context | 500K | 500K |
| Weights | Closed | Closed |
| Released | Aug 2026 | Sep 2026 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | Partial |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | ✓ |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Grok 4.6 → Grok 4.7
Reasoning cannot be disabled›
Reasoning cannot be disabled; choose low, medium, high or xhigh effort.
Penalties and stop return error›
presencePenalty, frequencyPenalty and stop cannot be used with reasoning models; requests containing them return an error.
Set prompt_cache_key›
Set prompt_cache_key or the Chat Completions x-grok-conv-id header to keep a conversation on the same server and improve cache hits.
Pass reasoning items back unchanged · Responses›
Responses always returns encrypted reasoning even without include; pass the reasoning items back unchanged in the next request.
10% premium · US regional endpoint›
The US regional endpoint keeps inference in the United States and adds a 10% token-price premium.
Grok 4.7 → Grok 4.6
Nothing known.