BlogLLM Benchmark Scores as Published, Side by Side
GLM-5.3 vs MiniMax-M3.1-Flash: Benchmark Scores Side by Side
Every test both GLM-5.3 by Z.ai and MiniMax-M3.1-Flash by MiniMax have a published number for: 11 tests in 8 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 GLM-5.3, MiniMax-M3.1-FlashData 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 | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| AutomationBench |
Agentic Coding
| Test | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| DeepSWE 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 | 73.7%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25; printed on 1 more pageHarness: mini-SWE-agent 2.4.5; Dataset revision: v1.1 |
| Terminal-Bench 4.0 | 37.9%DeepSeek, DeepSeek-V4.1-Flash model card, checked 2026-10-15; printed on 4 more pages | |
| NL2Repo-Bench | 58Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 3 more pagesEffort: max; Context: 1000000 |
ML Engineering
| Test | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| PostTrainBench v1.1 | 39.8%Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 1 more pageEffort: max; Harness: Claude Code 2.1.207; Context: 1000000 | 39.3%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25Harness: Claude Code 2.1.251; Dataset revision: v1.1 |
Long Context
| Test | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| LOCA-Bench (256k environment description) | 50.3%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25 | 60.3%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25 |
Multimodal Understanding
| Test | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| GDP.pdf | 11.2%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25; printed on 1 more page | 20.8%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25 |
Cybersecurity
| Test | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| CyberGym | 84.5%Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 2 more pagesEffort: max; Harness: Claude Code 2.1.207 | 82.2%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25; printed on 1 more pageHarness: Claude Code 2.1.263 |
General Knowledge and Reasoning
| Test | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| 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) | 43.4%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25Tools: none; Dataset revision: text-only subset (2,158 questions) |
Tool Use and Workflows
| Test | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| Agents' Last Exam (pass rate) | 28.5%Z.ai, GLM-5.3 model card, checked 2026-10-15; printed on 3 more pagesEffort: max; Harness: Claude Code; Context: 1000000 | 29.6%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25; printed on 1 more page |
| BankerToolBench | 74.5%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25 | 76.2%MiniMax, MiniMax-M3.1-Flash model card, checked 2026-10-25; printed on 1 more pageHarness: OpenCode 1.17.18 |
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 | GLM-5.3 | MiniMax-M3.1-Flash |
|---|---|---|
| Input | $1.4Z.ai, Z.AI API pricing, checked 2026-10-15 | $0.15MiniMax, MiniMax API pricing, checked 2026-10-19 |
| Cached input | $0.26Z.ai, Z.AI API pricing, checked 2026-10-15 | $0.015MiniMax, MiniMax API pricing, checked 2026-10-19 |
| Output | $4.4Z.ai, Z.AI API pricing, checked 2026-10-15 | $0.6MiniMax, MiniMax API pricing, checked 2026-10-19 |
| Cost per taskVals AI, USD, lower is better |
GLM-5.3 has 62 more and MiniMax-M3.1-Flash has 3 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.