BlogLLM Benchmark Scores as Published, Side by Side
Kimi K2 Thinking: Benchmark Scores as Published
Every result the catalog holds for Kimi K2 Thinking: 26 tests in 8 areas, from 3 pages. 22 printed by Moonshot AI itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer Moonshot AIReleased 2025-11-06Context window 256KInput TextOutput TextWeights Open (Modified MIT License)Data version 2026103001, as of 2026-10-30
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-bench Verified | 71.3% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 2 more pages | Effort: thinking; Harness: in-house harness derived from SWE-agent; Tools: enabled; bash and edit tools |
| SWE-bench Multilingual | 61.1% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Harness: in-house harness derived from SWE-agent; Tools: enabled; bash and edit tools |
| Terminal-Bench 2.0 | 35.7% | Z.ai, GLM-4.7 model cardchecked 2026-10-15 | - |
| Terminal-Bench (version not stated) | 47.1% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Harness: Terminus-2 with the provided JSON parser; Tools: enabled; simulated tools (JSON) |
| Terminal-Bench Hard | 30.6% | Z.ai, GLM-4.7 model cardchecked 2026-10-15 | - |
| Multi-SWE-bench | 41.9% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Harness: in-house harness derived from SWE-agent; Tools: enabled; bash and edit tools |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 84.5% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: none |
| AIME 2025 | 94.5% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: none |
| 100% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Harness: heavy mode: 8 parallel trajectories, aggregated; Tools: enabled; python | |
| 99.1% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: enabled; python | |
| SciCode | 44.8% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: none |
| HMMT February 2025 | 89.4% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: none |
| 95.1% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: enabled; python | |
| 97.5% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Harness: heavy mode: 8 parallel trajectories, aggregated; Tools: enabled; python | |
| IMO-AnswerBench | 78.6% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: none |
| HMMT November 2025 | 89.2% | Z.ai, GLM-4.7 model cardchecked 2026-10-15 | - |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 23.9% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: none |
| MMLU-Pro | 84.6% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: none |
| Humanity's Last Exam (with tools) | 51% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Harness: heavy mode: 8 parallel trajectories, aggregated; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| 44.9% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: enabled; search, code interpreter, web browsing; Context: 256000 | |
| MMLU-Redux | 94.4% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: none |
Code Generation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| LiveCodeBench v6 | 83.1% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: none |
| OJBench (C++) | 48.7% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: none |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| τ²-bench (domain not stated) | 74.3% | Z.ai, GLM-4.7 model cardchecked 2026-10-15 | - |
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 60.2% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| BrowseComp-ZH | 62.3% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: thinking; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| Seal-0 | 56.3% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| FRAMES | 87% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| FinSearchComp T3 | 47.4% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
Writing and Conversation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Longform Writing | 73.8 | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: none |
Healthcare
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| HealthBench | 58 | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: thinking; Tools: none |
More from Moonshot AI
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.