benchmark evidence
BenchLM Reasoning
BenchLM composite reasoning score (0-100).
activeupstream source ->
winner on BenchLM Reasoning
OpenAI: GPT-5.596.8
direct benchmark result, not a broad vertical composite | source row dated 2026-05-28
scored on 2026-05-28
latest mapped results | top 20
| # | Model | Score | Evidence | Tested |
|---|---|---|---|---|
| 1 | OpenAI: GPT-5.5 | 96.8 | model-only aggregator | 2026-05-28 |
| 2 | OpenAI: GPT-5.4 | 95.6 | model-only aggregator | 2026-05-28 |
| 3 | Anthropic: Claude Opus 4.6 | 88.4 | model-only aggregator | 2026-05-28 |
| 4 | OpenAI: GPT-5.2 | 83.8 | model-only aggregator | 2026-05-28 |
| 5 | Anthropic: Claude Sonnet 4.6 | 83.2 | model-only aggregator | 2026-05-28 |
| 6 | Anthropic: Claude Opus 4.5 | 70.1 | model-only aggregator | 2026-05-28 |
| 7 | OpenAI: GPT-5.1 | 67.3 | model-only aggregator | 2026-05-28 |
| 8 | Anthropic: Claude Sonnet 4.5 | 62.4 | model-only aggregator | 2026-05-28 |
| 9 | Z.ai: GLM 5 | 60.8 | model-only aggregator | 2026-05-28 |
| 10 | Google: Gemini 2.5 Pro | 59.4 | model-only aggregator | 2026-05-28 |
what this result means
BenchLM composite reasoning score (0-100).
This benchmark contributes direct public evidence. Read its scope before generalizing the result.
A win here is a win on BenchLM Reasoning. Broad task pages require independent corroboration before naming a general winner.
source record
category: reasoning
metric: score
matched models: 10
latest source date: 2026-05-28
direction: higher is better