Public benchmarkExternal source reviewReviewed by DanPublished Source checked Updated
Grok 4.7: separate release and public coding results
The report
Verdict
Published benchmark evidence
The 46.3% CursorBench, 37.6% Terminal-Bench and 71.0% DeepSWE claims come from Cursor's Grok partner presentation, not an independent DanManREAL coding trial.
On the tier list: The November 17 tier is Dan’s own ranking; public benchmark numbers are a separate evidence channel, not an automatic tier adjustment.
What the benchmarks report
- CursorBench 4.0: 46.3%. Grok 4.7; xHigh effort. Reported by Cursor/xAI partner release.
- Terminal-Bench 4.0: 37.6%. Grok 4.7; terminal agent coding. Reported by Cursor/xAI partner release.
- DeepSWE v1.1: 71.0%. Grok 4.7; high effort. Reported by Cursor/xAI partner release.
Original sources
Source 1 — CursorBench 4.0, Terminal-Bench 4.0, DeepSWE v1.1.
What this evidence tells a reader
The CursorBench 4.0 number is 46.3% at xHigh effort, while the DeepSWE v1.1 result names high effort. Different effort levels, test constructions and scoring methods make a single combined ranking misleading. Terminal-Bench 4.0 at 37.6% introduces yet another category: extended terminal work.
Cursor helped develop and presents this model, so its evaluation is valuable product-publisher evidence but should not be advertised as an independent third-party win. Cursor's published comparison also separates Grok 4.7 from older Grok 4.6; transferring one release's benchmark to the other would mislead a reader.
If choosing this model for long coding jobs, request the exact model version and effort setting. Then verify whether actual diffs are acceptable and whether the assistant recovers from failed commands. The scorecard cannot substitute for that implementation check.
How to use these results
- Use partner attribution whenever citing CursorBench claims from this page.
- Match the 4.7 release and effort level before comparing to another model.
- Judge the fit of a trial by accepted patches, rollback safety and completed work, not marketing claims about long-horizon coding.
Limits of this report
- The cited model partner also publishes the main evaluation; independence of these scores is not established.
- No corresponding Grok 4.7 local terminal run or cross-version controlled comparison is asserted here.
Follow the lab
New test reports and tier-list changes get announced in my Discord's #news channel. The full videos are on YouTube.
Receipts
DanManREAL · AI lab
Test receipt
Grok 4.7
- CursorBench 4.0
- 46.3%Cursor/xAI partner release (model partner publication). Grok 4.7; xHigh effort. Not independently rerun by DanManREAL.
- Terminal-Bench 4.0
- 37.6%Cursor/xAI partner release (model partner publication). Grok 4.7; terminal agent coding. Not independently rerun by DanManREAL.
- DeepSWE v1.1
- 71.0%Cursor/xAI partner release (model partner publication). Grok 4.7; high effort. Not independently rerun by DanManREAL.
- Source review
- November 17, 2026Date these source claims were reviewed; not a claim about when the published benchmark was executed.
Verdict
Published benchmark evidence
Tier list The November 17 tier is Dan’s own ranking; public benchmark numbers are a separate evidence channel, not an automatic tier adjustment.
danmanreal.com/tests/grok-4-7-public-evidence
Models covered by external benchmarks
On the tier list's bench
- Grok 4.7 · 1 receipt on file
Related
Independent testing. Sponsorships are always disclosed, and verdicts are never for sale.
Disagree? Bring your receiptsAll reports
- One inbox for my AI coding agents: triage took 34% less time, resuming work did not improveHead-to-headNov 20
- Mythos 5.2: research and security benchmark source reviewPublic benchmarkNov 17
- Fable 5.2: agentic coding benchmark source reviewPublic benchmarkNov 17
- Claude Opus 5.5: public coding benchmark receiptsPublic benchmarkNov 17
- Astra 6.2: new model and public task-evaluation evidencePublic benchmarkNov 17
- Sol 6.2: new model and public coding benchmark evidencePublic benchmarkNov 17
- Gemini 4 Carbon: benchmark source and usage profilePublic benchmarkNov 17
- Qwen 4: Qwen4-Max launch and benchmark sourcesPublic benchmarkNov 17
- Kimi K3.1 Pro: public coding benchmark receiptsPublic benchmarkNov 17
- GLM-5.3: public coding benchmark receiptsPublic benchmarkNov 17
- Gemini 4 Argon: public coding benchmark receiptsPublic benchmarkNov 17
- Claude Sonnet 5.5: public coding benchmark receiptsPublic benchmarkNov 17
- DeepSeek V4.1 Flash: public coding benchmark receiptsPublic benchmarkNov 17
- Muse Wonder Full 1.0: public coding benchmark receiptsPublic benchmarkNov 17
- Grok 4.7: separate release and public coding resultsPublic benchmarkNov 17
- ai-design-check on two of my FABLE screens: 10 warnings, 4 of them false alarmsField testNov 2
- Vercel Agent Desktop used a sample Trip app in 6 actions without moving the mouseField testNov 1
- Cerberus Guard on sample data: one crawler name, two verdicts, and a block earned by behaviorField testOct 31
- GPT-6.1 Terra vs Opus 5.5: three real projects, and the designs I picked blindHead-to-headOct 29
- Steps → Reduced: opening a terminal agent went from 98 seconds to 11AuditOct 28
- One ChatGPT chat directed Opus and Codex at once through DMR_REMOTEField testOct 28
- Claude Duet ran my channel for a week and caught 269 fake AI crawlersField testOct 27
- Kimi K3.1 Pro built my animation family, after a quota stop and 21 minutes of failed requestsField testOct 24
- ChatGPT + Codex built 33 new DMR_REMOTE toolsField testOct 23
- Gemini 4 Argon reviewed a finished episode: 14 clips, one flag, one inconsistencyField testOct 21
- Qwen 3.8 Flash Next on my RTX 5080: 81 tokens a second with 202K tokens readAuditOct 21
- Does an extra AI review step pay off? A small DeepSeek trial said noHead-to-headOct 19
- My real DeepSeek bill: 12.31 billion tokens for $122.28Cost receiptOct 17
- Gemini 4 Argon kept hanging in Gemini CLI. The cause was a 60-second timeoutIncidentOct 14
- I asked 10 AI coding agents for their biggest weaknessSelf-reportOct 14
- Jarvis 27B on my own PC: 30 tokens a second, and it still recalls at 251KAuditOct 10
- My agents' first game benchmark was wrong. Here's how it was caughtIncidentOct 10
- Pixal3D + Blender turned my 2D art into animated 3D charactersField testSep 22
- DeepSeek Flash answered 200 OK and sent nothing for two hoursIncidentSep 14
- Grok, Codex and DeepSeek played WorldOS for Rank 1. Grok won with 98Head-to-headAug 20
- DeepSeek V4 Flash audited 18 of my real coding sessions. Two said 'done' too earlyAuditAug 4