BlogLLM Benchmark Scores as Published, Side by Side
Grok 4: Benchmark Scores as Published
Every result the catalog holds for Grok 4: 6 tests in 2 areas, from 1 page. Each was printed by the company named beside it. Each number is as printed; nothing is ranked or rescored.
Developer xAIReleased 2025-07-09Data version 2026103001, as of 2026-10-30
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 87.5% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: none |
| AIME 2025 | 98.8% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: enabled; python |
| 91.7% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: none | |
| 100% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Harness: heavy mode: 8 parallel trajectories, aggregated; Tools: enabled; python | |
| HMMT February 2025 | 96.7% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Harness: heavy mode: 8 parallel trajectories, aggregated; Tools: enabled; python |
| 93.9% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: enabled; python | |
| 90% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: none | |
| IMO-AnswerBench | 73.1% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: none |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 25.4% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: none |
| Humanity's Last Exam (with tools) | 41% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| 50.7% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Harness: heavy mode: 8 parallel trajectories, aggregated; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
More from xAI
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.