live
weekly refresh
basedagi.org
benchmark evidence

BenchLM Coding

BenchLM composite coding score (0-100).

winner on BenchLM Coding
direct benchmark result, not a broad vertical composite | source row dated 2026-07-09
scored on 2026-07-09
latest mapped results | top 20
#ModelScoreEvidenceTested
1OpenAI: GPT-5.4
Openai
86.4
model-only
aggregator
2026-07-09
2Anthropic: Claude Opus 4.6
Anthropic
82.4
model-only
aggregator
2026-07-09
3DeepSeek: DeepSeek V4 Flash
Deepseek
81.8
model-only
aggregator
2026-07-09
4Anthropic: Claude Sonnet 4.6
Anthropic
75.7
model-only
aggregator
2026-07-09
5Anthropic: Claude Sonnet 4.5
Anthropic
74.4
model-only
aggregator
2026-07-09
6OpenAI: GPT-5.2
Openai
73.7
model-only
aggregator
2026-07-09
7Anthropic: Claude Opus 4.5
Anthropic
73.0
model-only
aggregator
2026-07-09
8OpenAI: GPT-5.5
Openai
72.5
model-only
aggregator
2026-07-09
9Z.ai: GLM 5
Z Ai
71.9
model-only
aggregator
2026-07-09
10OpenAI: GPT-5.1
Openai
69.5
model-only
aggregator
2026-07-09
11OpenAI: GPT-5.4 Mini
Openai
68.5
model-only
aggregator
2026-07-09
12DeepSeek: DeepSeek V4 Pro
Deepseek
68.1
model-only
aggregator
2026-07-09
13OpenAI: GPT-4.1
Openai
61.5
model-only
aggregator
2026-07-09
14Google: Gemini 2.5 Pro
Google
46.1
model-only
aggregator
2026-07-09
what this result means

BenchLM composite coding score (0-100).

This benchmark contributes direct public evidence. Read its scope before generalizing the result.

A win here is a win on BenchLM Coding. Broad task pages require independent corroboration before naming a general winner.

source record
category: coding
metric: score
matched models: 14
latest source date: 2026-07-09
direction: higher is better
inspect upstream source ->