Skip to content
Sir CompareComparisons with evidenceAsk Sir CompareAsk
United Kingdom · GBP
Country and pricesUnited KingdomGBPUnited StatesUSDItalyEURGermanyEURSpainEURFranceEUR
Language: English
ITENDEFRES

Comparison · AI Models & LLMs

GLM-5 vs Claude Opus 4.6

The same method as our rankings: scores from the engine, sourced data, verified prices.

Details: Claude Opus 4.6 →Details: GLM-5 →

★ Winner
Anthropic Claude Opus 4.6 — drawn card, not a photograph of the product

Claude Opus 4.6

Anthropic

Full sheet →

81.9/100

SWE-bench accuracy
75.6%
Context window
1M tokens
Input cost per 1M tokens
$5.00
Output cost per 1M tokens
$25.00
Z.ai GLM-5 — drawn card, not a photograph of the product

GLM-5

Z.ai

Full sheet →

71.3/100

SWE-bench accuracy
72.8%
Context window
200k tokens
Maximum output
128k tokens
Input cost per 1M tokens
$1.00

The verdicttake the Claude Opus 4.6, it leads in 2 of 4 areas — above all on context window.

The other one stays ahead on Output cost per 1M tokens.

  • Claude Opus 4.6 better for: context window, SWE-bench Verified (resolved %)
  • GLM-5 better for: Output cost per 1M tokens, Input cost per 1M tokens

Basis: 9 measured differences across 9 compared characteristics. The scores come from the ranking engine.

The numbers side by side

Each bar is the product's verified figure; on every row the longer bar is the higher of the two values.

  • SWE-bench accuracyhigher is better
    Claude Opus 4.675.6%
    GLM-572.8%
  • Context windowhigher is better
    Claude Opus 4.61M tokens
    GLM-5200k tokens
  • Input cost per 1M tokenslower is better
    Claude Opus 4.6$5.00
    GLM-5$1.00
  • Cached input cost per 1M tokenslower is better
    Claude Opus 4.6$0.50
    GLM-5$0.20
  • Output cost per 1M tokenslower is better
    Claude Opus 4.6$25.00
    GLM-5$3.20

Who wins where

  • For codingClaude Opus 4.6SWE-bench accuracy
  • For long documentsClaude Opus 4.6Context window
  • Cheaper to readGLM-5Input cost per 1M tokens
  • Cheaper to writeGLM-5Output cost per 1M tokens

GLM-5: open weights (downloadable)

Data sheet side by side

Performance

SWE-bench accuracy✓75.6%3.8% more72.8%

Memory and answer length

Context window✓1M tokens400% more200k tokens
Maximum outputNot stated128k tokens

API prices

Input cost per 1M tokens$5.00✓$1.0080% less
Cached input cost per 1M tokens$0.50✓$0.2060% less
Output cost per 1M tokens$25.00✓$3.2087% less

Size and openness

ParametersNot stated744 billion
Active parameters per tokenNot stated40 billion
Open weights (downloadable)Not statedYes

What a month costs

Change the volumes: the cost is worked out from the verified list prices above.

Claude Opus 4.6$22.50per month · $270 per year
GLM-5$3.60per month · $43.20 per yearThe cheaper one — you save $18.90 per month

The scenarios are only starting points to edit, not data about your use. List prices without discounts, time-of-day rates or taxes.

Questions and answers

For coding: Claude Opus 4.6 or GLM-5?

Claude Opus 4.6: 75.6% against 72.8% (SWE-bench accuracy).

For long documents: Claude Opus 4.6 or GLM-5?

Claude Opus 4.6: 1M tokens against 200k tokens (Context window).

Cheaper to read: Claude Opus 4.6 or GLM-5?

GLM-5: $1.00 against $5.00 (Input cost per 1M tokens).

Cheaper to write: Claude Opus 4.6 or GLM-5?

GLM-5: $3.20 against $25.00 (Output cost per 1M tokens).

Area by area

The same yardstick for both: the engine's score out of 100 on every criterion of the profile.

  • SWE-bench Verified (resolved %)
    10096
  • Context window
    10020
  • Output cost per 1M tokens
    969
  • Input cost per 1M tokens
    1260
Compare other products

Your assistant

Open full page →