1. Chart data
  2. github:zai-org/GLM-4
  3. Model Performance Comparison Across Benchmarks
github:zai-org/GLM-4Benchmark bar chart · 30 points · DeepSeek R1

Model Performance Comparison Across Benchmarks

The numbers behind a chart github:zai-org/GLM-4 published, redrawn natively. Check the original, study the data, then make your own.

Estimated from the image 93%DeepSeek R1All DeepSeek modelsAll chart data
Chart by github:zai-org/GLM-4. Source: github.com

Our replica

Look-alike score 9.0 of 10
The original, github:zai-org/GLM-4
Our replica. Look-alike score 9.0 of 10Swipe sideways to read the labels.

Drawn by us in your browser from the numbers below, in the same layout, palette and label style as the original. Values marked with ~ in the table are estimates. A replica needs 7 of 10 from our judges to be shown.

The numbers

First 10 of 30 points are free
GroupSeriesValueSourceConf.
AIME24 数学推理Z1-9B-041476.4Estimate0.97
AIME24 数学推理DeepSeek-R1-Distill-Qwen-7B55.5Estimate0.97
AIME24 数学推理DeepSeek-R1-Distill-Qwen-14B69.7Estimate0.50
AIME25 数学推理Z1-9B-041456.6Estimate0.97
AIME25 数学推理DeepSeek-R1-Distill-Qwen-7B40.4Estimate0.97
AIME25 数学推理DeepSeek-R1-Distill-Qwen-14B48.9Estimate0.97
Omni-MATH 数学推理Z1-9B-041464Estimate0.97
Omni-MATH 数学推理DeepSeek-R1-Distill-Qwen-7B49Estimate0.97
Omni-MATH 数学推理DeepSeek-R1-Distill-Qwen-14B56.1Estimate0.97
LCB (2408-2501) 代码推理Z1-9B-041451.8Estimate0.97

Max data, free with your email for now

See all 30 points and download the files

The full table, the JSON, a CSV and the agent skill file for this chart. These are Max in the plan; while this is a prototype your email unlocks them.

Unlocks all 35 giants. One email a week, leave any time. What you get · Log in

LCB (2408-2501) 代码推理DeepSeek-R1-Distill-Qwen-7B37.6Estimate0.97
LCB (2408-2501) 代码推理DeepSeek-R1-Distill-Qwen-14B53.1Estimate0.97
SysBench (ISR) 指令遵循Z1-9B-041477.3Estimate0.97
SysBench (ISR) 指令遵循DeepSeek-R1-Distill-Qwen-7B25.9Estimate0.97
SysBench (ISR) 指令遵循DeepSeek-R1-Distill-Qwen-14B53.7Estimate0.97
IFEval 指令遵循Z1-9B-041481.9Estimate0.97
IFEval 指令遵循DeepSeek-R1-Distill-Qwen-7B58.4Estimate0.97
IFEval 指令遵循DeepSeek-R1-Distill-Qwen-14B72.3Estimate0.97
ArenaHard 通用问答Z1-9B-041467.4Estimate0.97
ArenaHard 通用问答DeepSeek-R1-Distill-Qwen-7B31.3Estimate0.30
ArenaHard 通用问答DeepSeek-R1-Distill-Qwen-14B50.4Estimate0.97
BFCL (v3) 综合工具调用Z1-9B-041466.6Estimate0.97
BFCL (v3) 综合工具调用DeepSeek-R1-Distill-Qwen-7B24Estimate0.97
BFCL (v3) 综合工具调用DeepSeek-R1-Distill-Qwen-14B50.2Estimate0.97
SuperGPQA 科学Z1-9B-041443.6Estimate0.97
SuperGPQA 科学DeepSeek-R1-Distill-Qwen-7B7.2Estimate0.97
SuperGPQA 科学DeepSeek-R1-Distill-Qwen-14B40.3Estimate0.97
GPQA 科学Z1-9B-041458.5Estimate0.97
GPQA 科学DeepSeek-R1-Distill-Qwen-7B49.1Estimate0.97
GPQA 科学DeepSeek-R1-Distill-Qwen-14B59.1Estimate0.97
Download this chart's data

Max subscribers get the same data by API. API docs

JSONAgent skill file

API · Max

Use this chart via API

On the Max plan you can pull this chart's numbers with the provenance of each one, as JSON or CSV, with its redraw spec, or call it from an agent over MCP. Chart id e689c3f2054d.

See the Max planAPI docs
PathReturnsPlan
GET/api/v1/charts/e689c3f2054dTitle, publisher, date, form, point count; with a Max key, the data linkOpen
GET/api/v1/charts/e689c3f2054d/dataEvery point with its provenance and confidence, plus the redraw spec. Add ?format=csv for CSVMax
GET/api/v1/charts/e689c3f2054d/render?view=1The draw-only view this page usesOpen
POST/mcpTool get_chart_data with this id, for Claude, Cursor and other agentsMax
curlMax
curl -s https://viral-chart.com/api/v1/charts/e689c3f2054d/data \
  -H "Authorization: Bearer $VIRALCHART_KEY"
Claude CodeMax
claude mcp add --transport http viralchart https://viral-chart.com/mcp \
  --header "Authorization: Bearer $VIRALCHART_KEY"

60 requests a minute per key; over that you get 429 with Retry-After. Up to 5 keys per account, made on your account page and shown once. Without Max, the data route answers 402 with a link to the plans. The numbers describe a chart its publisher owns, and every file carries a license note.

Make one like this

This chart's shape, your numbers.

Live preview. Everything stays on your device.

Pick the form

Start from our replica of this chart, or from a simplified form. Either way your numbers start from this chart's real data.

More data from DeepSeek

All chart data
Partly publisher's numbersReplica 8.1Accuracy / Percentile (%)DeepSeek · Benchmark bar chart · 30 pointsEstimated from the imageGlobal KV Cache Per Token (Bytes)DeepSeek · Bar chart · 4 pointsEstimated from the imageReplica 6.3AI Model Benchmark Performance Comparisonqwen · Benchmark bar chart · 25 pointsPartly publisher's numbersReplica 8.3Accuracy / Percentile (%)DeepSeek · Benchmark bar chart · 30 pointsPartly publisher's numbersAccuracy / Percentile (%)DeepSeek · Benchmark bar chart · 36 pointsEstimated from the imageReplica 9.0AI Model Benchmarks ComparisonDeepSeek · Benchmark bar chart · 20 points

Questions people ask

How accurate are these numbers?

They are estimated from the image by two independent passes, with confidence 0.93 on a 0 to 1 scale. Printed numbers read well; values read off an axis can be a few percent off.

Can I reuse this data?

The numbers describe a chart that github:zai-org/GLM-4 published. The image, layout and styling belong to them. Use the numbers with a credit to the publisher, and make your picture from your own data.

How do I get the file?

The JSON, CSV and agent skill file are under the email gate on this page. In this prototype they are static files at /api/charts/e689c3f2054d.json and /api/charts/e689c3f2054d.skill.md.