BlogLLM Benchmark Scores as Published, Side by Side

Apex Bench: Results as Published

Every result the catalog holds for Apex Bench: 6 models from 5 developers, read from 1 page. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.

Area Knowledge WorkData version 2026103001, as of 2026-10-30

The DeepSeek-V4.1-Flash-1017 model card prints "Apex Bench v1.1" as one row of its comparison table and does not say who maintains the test, what the tasks are, or how a result is scored. Mercor publishes a version 1.1 of APEX-Agents (September 2026), and the catalog does not establish that the two are the same test: no value on this card matches a value printed elsewhere for APEX-Agents, APEX-SWE or MathArena Apex.

Not checked against a maintainer's page.

Apex Bench v1.1

ModelValuePrinted bySetup as printed
GPT-5.6 SolOpenAI, released 2026-07-0954.3%DeepSeek, DeepSeek-V4.1-Flash-1017 model cardchecked 2026-10-18-
Claude Opus 5Anthropic, released 2026-07-2465.8%DeepSeek, DeepSeek-V4.1-Flash-1017 model cardchecked 2026-10-18-
DeepSeek-V4.1-Flash-1017DeepSeek, released 2026-10-1772.1%DeepSeek, DeepSeek-V4.1-Flash-1017 model cardchecked 2026-10-18-
DeepSeek-V4.1-FlashDeepSeek, released 2026-09-1064.6%DeepSeek, DeepSeek-V4.1-Flash-1017 model cardchecked 2026-10-18-
Kimi K3Moonshot AI, released 2026-07-1647.8%DeepSeek, DeepSeek-V4.1-Flash-1017 model cardchecked 2026-10-18-
GLM-5.3Z.ai, released 2026-08-1856.6%DeepSeek, DeepSeek-V4.1-Flash-1017 model cardchecked 2026-10-18-

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. 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.