Navigate
Advertise here

DeepSeek V3.2 (medium) vs Ember-1 (high)

The average score is effectively tied at 7.0 vs 7.1. DeepSeek V3.2 (medium) has the lower benchmark cost at $0.081 vs $1.798. Ember-1 (high) is faster at 56.77s vs 68.43s, with pass rates of 65.2% vs 71.2%.

Last updated at: 2026-09-28

Compared models

Rank
#162
Total Output Tokens
128,787
Response Time (avg)
68.43s
Total Cost
$0.081
Rank
#158
Total Output Tokens
110,125
Response Time (avg)
56.77s
Total Cost
$1.798
Recommended model DeepSeek V3.2 (medium)

It has the best score here (7.0), while costing about 22.3x less than Ember-1 (high).

Detailed comparison

Metric DeepSeek V3.2 DeepSeek V3.2 medium Release: 2025-12-01 Ember-1 Ember-1 high Release: 2026-09-28
Score 7.0 7.1
Rank #162 #158
Reliability 10.0 8.0
Consistency 7.4 7.3
Attempts 66/66 66/66
Tests Correct
Attempt pass rate 65.2% 71.2%
Flaky tests 7 7
Total Runs 66 66
Cost per result 0.672 14.981
Total Cost $0.081 $1.798
Input Price $0.280 / 1M $3.000 / 1M
Output Price $0.420 / 1M $15.000 / 1M
Total Input Tokens 101,056 48,577
Output Tokens 11,832 7,345
Reasoning Tokens 116,955 102,780
Response Time (avg) 68.43s 56.77s
Response Time (max) 376.10s 255.06s
Response Time (total) 1505.53s 1248.88s
Parameters 671B total (37B active) ~2.8T total (~104B active)
Availability Open source Closed

Model generation showcase

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#162 DeepSeek V3.2

medium
Cost
$0.001
Time
53.6s
Tokens
1,932 tok

#158 Ember-1

high
Cost
$0.038
Time
55.1s
Tokens
2,671 tok

Top Models by Score

Score vs Total Cost

Response Time (avg)

Score vs Response Time (avg)

Total Output Tokens

Score vs Total Output Tokens

Category Breakdown

Coding Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V3.2 6.0 7.2 55.6% 1 248.68s 5,717 649 52,014
Ember-1 8.2 7.2 88.9% 1 72.23s 7,144 1,281 23,093

Quick Compare

Switch Comparison Pair