vakati-tech.ai

// head_to_head

Kimi K3 vs GPT-5.6 Terra

Published LiveBench scores across all seven categories, live list pricing, context windows, and the measured cost of a point of capability — for both models, side by side.

OpenRouter + LiveBenchAll comparisonsFull leaderboard

GPT-5.6 Terra is the better model, at roughly the same price.

GPT-5.6 Terra leads by 2.2 points overall and the two list within about 10% of each other, so the cheaper-but-weaker trade-off does not apply. Price parity plus a score gap usually makes this an easy call.

moonshotai

Kimi K3

Blended / 1M
$4.14
Context
1.0M
Released
Jul 16, 2026
Overall score
65.0
reasoningtool callingimage inputvideo inputprompt caching

openai

GPT-5.6 Terra

Blended / 1M
$4.50
Context
1.1M
Released
Jul 9, 2026
Overall score
67.2
reasoningtool callingfile inputimage inputprompt caching

Specs and pricing

MetricKimi K3GPT-5.6 Terra
LiveBench overall

Weighted mean of the seven LiveBench category scores, 0–100. Higher is better.

65.067.2win
Cost per point

Measured benchmark spend divided by overall score — dollars per point of capability.

$0.0217$0.0209
Blended price / 1M

3:1 input:output mix, the usual shape of production traffic.

$4.14$4.50
Input price / 1M$0.192win$2.00
Output price / 1M$16.00$12.00win
Cached input / 1M

Price of an input token served from the prompt cache, where the provider publishes one.

$0.182$0.200
Context window1.0M1.1M
Max output tokens944Kwin128K

Benchmarks by category

An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — Kimi K3 on top, GPT-5.6 Terra below, both out of 100.

Long-horizon agentic
36.0
25.9
Coding
53.1
84.2
Reasoningtoo close to call
90.7
90.6
Browser
75.2
70.3
Data analysistoo close to call
78.7
79.3
Researchtoo close to call
79.8
79.7
Documents & design
72.5
70.9

What each one costs to run

Per-token prices are hard to feel. These are monthly list costs for both models across five workload shapes, using each provider's published cached-input rate where there is one.

WorkloadKimi K3GPT-5.6 Terra
Support chatbot

1.2K in / 400 out × 200K requests

$1,325/mo$1,310/mo
RAG assistant

8K in / 600 out × 100K requests

$1,110/mo$1,600/mo
Coding agent

40K in / 4K out × 20K requests

$1,428/mo$1,552/mo
Document extraction

20K in / 1.5K out × 50K requests

$1,392/mo$2,810/mo
Bulk classification

500 in / 20 out × 5M requests

$2,075/mo$5,300/mo
Run these two through the cost calculator

Which should you pick?

Quality matters more than the bill

GPT-5.6 Terra

Highest overall LiveBench score of the two at 67.2.

The workload is coding or agentic work

Kimi K3

Leads on long-horizon agentic — 36.0 against 25.9.

Kimi K3 vs GPT-5.6 Terra FAQ

Which is better, Kimi K3 or GPT-5.6 Terra?

GPT-5.6 Terra is the better model, at roughly the same price. GPT-5.6 Terra leads by 2.2 points overall and the two list within about 10% of each other, so the cheaper-but-weaker trade-off does not apply. Price parity plus a score gap usually makes this an easy call.

Is Kimi K3 cheaper than GPT-5.6 Terra?

They cost about the same. Both land near $4.14 per million tokens on a 3:1 input:output blend, so price is unlikely to be the deciding factor between them.

Kimi K3 vs GPT-5.6 Terra: which scores higher on benchmarks?

Kimi K3 scores 65.0 and GPT-5.6 Terra scores 67.2 overall on LiveBench, the weighted mean of its seven categories. That is a 2.2-point lead for GPT-5.6 Terra. Category scores differ from the overall figure — a model can lead on reasoning and trail on coding, which the per-category table above breaks out.

Which gives better value for money, Kimi K3 or GPT-5.6 Terra?

They are close. Kimi K3 costs $0.0217 per point of overall capability and GPT-5.6 Terra costs $0.0209, a difference small enough that workload shape will matter more than the rate.

Does Kimi K3 or GPT-5.6 Terra have a bigger context window?

They are effectively the same — 1.0M for Kimi K3 and 1.1M for GPT-5.6 Terra.

Do Kimi K3 and GPT-5.6 Terra support prompt caching?

Both publish a cached-input rate: $0.182 per million for Kimi K3 and $0.200 for GPT-5.6 Terra, against full input rates of $0.192 and $2.00. On a workload with a long stable prefix — a system prompt, a tool schema, a retrieved corpus — that changes the economics more than the headline price does.

Related comparisons

How these numbers are produced

  • Price — provider list price from OpenRouter, refreshed every 15 minutes. “Blended” is a 3:1 input:output mix.
  • Scores — LiveBench release 2026-11-06, using their own category map. Each model shows its strongest published run. A blank means “not evaluated”, never “bad”.
  • Cost per point — the measured dollars LiveBench spent on the run, divided by the score it earned.
  • “Win” — awarded only past a threshold: one full point on a benchmark score, 10% on a price, 25% on a context window. Anything tighter reports as a tie, because effort settings alone move a LiveBench score by more than that.

Published benchmarks rank models on someone else's tasks. Before committing, see LLM & agent evaluation for building an eval on your own.