BlogLLM Benchmark Scores as Published, Side by Side
Kimi K3.1 Pro vs GLM-5.3: Benchmark Scores Side by Side
Every test both Kimi K3.1 Pro by Moonshot AI and GLM-5.3 by Z.ai have a published number for: 33 tests in 10 areas. Each number is as printed, with the page it comes from and the setup it was run under; nothing is ranked or rescored, and two numbers under different setups are two results, not one comparison.
Models Kimi K3.1 Pro, GLM-5.3Data version 2026103001, as of 2026-10-30
Model
What each developer prints on its own model page; a dash means it prints none.
Knowledge Work
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| AutomationBench | ||
| APEX-Agents | ||
| OfficeQA Pro | ||
| GDPval-AA v2.1 | ||
| AA-Briefcase v1.1 | ||
| Business Operation (Internal) | ||
| FrontierFinance | 58.6%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Finance Agent v2 | 57.5%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Finance Agent v2 |
| OfficeQA Pro v2 | ||
| DECK-Arena (Internal, Elo) |
Agentic Coding
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| DeepSWE v1.1 | 75.5%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Claude Code; Dataset revision: v1.1 | 66.9%Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 6 more pagesEffort: max; Harness: mini-swe-agent; Context: 400000 |
| Terminal-Bench 4.0 | 37.9%DeepSeek, DeepSeek-V4.1-Flash model card, checked 2026-10-15; printed on 4 more pages | |
| ProgramBench Almost@1 | ||
| SWE-Marathon | 51.7%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Claude Code; Dataset revision: v1.1 | |
| Kimi Code Bench 2.0 | ||
| Kimi Code Bench 3.0 Long-horizon |
ML Engineering
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| PostTrainBench v1.2 | 36.5%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Claude Code; Dataset revision: v1.2 |
Science and Math
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| GPQA Diamond | 88.1%DeepSeek, DeepSeek-V4.1-Flash model card, checked 2026-10-15; printed on 2 more pages | |
| Terminal-Bench Science 0.1 | ||
| BioMysteryBench |
Multimodal Understanding
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| GDP.pdf | 11.2%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25; printed on 1 more page |
Cybersecurity
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| ExploitBench | 74.7%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: ExploitBench official harness; Attempt budget: 300 turns; pass@5 | 54.4%Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 1 more pageEffort: max; Harness: Claude Code 2.1.207 |
| ExploitGym | 20.7%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Kimi Code; Attempt budget: 6-hour wall-clock budget per task | 15%DeepSeek, DeepSeek-V4.1-Flash model card, checked 2026-10-15; printed on 2 more pages |
| xdev-bench-hard, problems solved (pass@3) | 81%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Harness: Kimi Code; Attempt budget: pass@3 | 57.1%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Harness: Claude Code; Attempt budget: pass@3 |
| xdev-bench-hard, average success rate | 64.5%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Harness: Kimi Code; Attempt budget: pass@3 | 34.9%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Harness: Claude Code; Attempt budget: pass@3 |
General Knowledge and Reasoning
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| Humanity's Last Exam (no tools) | 42%DeepSeek, DeepSeek-V4.1-Flash model card, checked 2026-10-15; printed on 3 more pagesDataset revision: text-only subset of HLE (dagger on the card) | |
| Humanity's Last Exam (with tools) | 62.5%Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 3 more pagesEffort: max; Context: 300000 | |
| Humanity's Last Exam Diamond (no tools) | 16.4%Meta, Meta Muse Spark Open 816B-A42B model card, checked 2026-10-19; printed on 1 more pageEffort: max; Tools: none | |
| Kimi Hallu Bench Hard w/ tools (lower is better)lower is better | ||
| Kimi Hallu Bench w/o tools (lower is better)lower is better |
Tool Use and Workflows
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| Agents' Last Exam (pass rate) | 43.2%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Kimi Code; Dataset revision: v1.1 | 28.5%Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 3 more pagesEffort: max; Harness: Claude Code; Context: 1000000 |
| Agents' Last Exam (CLI) | 44.2%Moonshot AI, Introducing Kimi K3.1 Pro, checked 2026-10-24Effort: max; Harness: Kimi Code; Dataset revision: v1.1 |
Instruction Following
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| HANDBOOK.md |
Engineering and Design
| Test | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| Design Bench (Internal, Elo) |
Price
API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked; a dash means the developer lists that price in another unit or not at all. Cost per task is measured by Vals AI.
| Price | Kimi K3.1 Pro | GLM-5.3 |
|---|---|---|
| Input | $2Moonshot AI, Kimi API pricing (international platform), checked 2026-10-24 | $1.4Z.ai, Z.AI API pricing, checked 2026-10-15 |
| Cached input | $0.2Moonshot AI, Kimi API pricing (international platform), checked 2026-10-24 | $0.26Z.ai, Z.AI API pricing, checked 2026-10-15 |
| Cache write | - | |
| Cache write, 1 hour | $4Moonshot AI, Kimi API pricing (international platform), checked 2026-10-24 | - |
| Output | $10Moonshot AI, Kimi API pricing (international platform), checked 2026-10-24 | $4.4Z.ai, Z.AI API pricing, checked 2026-10-15 |
| Cost per taskVals AI, USD, lower is better |
Kimi K3.1 Pro has 18 more and GLM-5.3 has 40 more: tests only one of them has a number for, listed on its model page.
More side by side
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. A pair has a page when both models are their developers' flagships and both have a number for at least 10 of the same tests. Where one page printed a 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.