// head_to_head

Kimi K2.7 Code vs Qwen3.5 397B A17B

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

Kimi K2.7 Code and Qwen3.5 397B A17B are priced within ~10% of each other.

Qwen3.5 397B A17B does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.

moonshotai

Kimi K2.7 Code

Blended / 1M
$1.35
Context
262K
Released
Jun 12, 2026
Overall score
68.4
reasoningtool callingimage inputprompt caching

qwen

Qwen3.5 397B A17B

Blended / 1M
$1.29
Context
262K
Released
Feb 16, 2026
Overall score
Not evaluated
reasoningtool callingimage inputvideo inputprompt caching

Specs and pricing

MetricKimi K2.7 CodeQwen3.5 397B A17B
LiveBench overall

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

68.4
Cost per point

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

$0.0545
Blended price / 1M

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

$1.35$1.29
Input price / 1M$0.706$0.550win
Output price / 1M$3.30$3.50
Cached input / 1M

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

$0.180win$0.225
Context window262K262K
Max output tokens236K236K

Benchmarks by category

An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — Kimi K2.7 Code on top, Qwen3.5 397B A17B below, both out of 100.

Agentic coding
45.7
Coding
74.0
Reasoning
82.8
Mathematics
79.6
Data analysis
62.7
Language
77.9
Instruction following
56.3

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 K2.7 CodeQwen3.5 397B A17B
Support chatbot

1.2K in / 400 out × 200K requests

$395.60/mo$388.60/mo
RAG assistant

8K in / 600 out × 100K requests

$552.48/mo$520.00/mo
Coding agent

40K in / 4K out × 20K requests

$534.29/mo$538.00/mo
Document extraction

20K in / 1.5K out × 50K requests

$927.39/mo$796.25/mo
Bulk classification

500 in / 20 out × 5M requests

$1,832/mo$1,563/mo
Run these two through the cost calculator

Kimi K2.7 Code vs Qwen3.5 397B A17B FAQ

Which is better, Kimi K2.7 Code or Qwen3.5 397B A17B?

Kimi K2.7 Code and Qwen3.5 397B A17B are priced within ~10% of each other. Qwen3.5 397B A17B does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.

Is Kimi K2.7 Code cheaper than Qwen3.5 397B A17B?

They cost about the same. Both land near $1.35 per million tokens on a 3:1 input:output blend, so price is unlikely to be the deciding factor between them.

Does Kimi K2.7 Code or Qwen3.5 397B A17B have a bigger context window?

They are effectively the same — 262K for Kimi K2.7 Code and 262K for Qwen3.5 397B A17B.

Do Kimi K2.7 Code and Qwen3.5 397B A17B support prompt caching?

Both publish a cached-input rate: $0.180 per million for Kimi K2.7 Code and $0.225 for Qwen3.5 397B A17B, against full input rates of $0.706 and $0.550. 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.
  • ScoresLiveBench release 2026-06-25, 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.