Gemini 3.6 Flash vs Kimi K3
Which one, when — curated verdict, not a benchmark table
Both share a 1,048,576-token context window, so the split is price, output length, and inputs: gemini-3.6-flash runs $1.5 per million input and $7.5 output, exactly half kimi-k3's $3 and $15, and it is the only one that accepts audio (billed at $5 per million audio input tokens, $0.5 cached). Pick gemini-3.6-flash for cheaper high-volume text, image, video or audio work, with cache reads at $0.15 versus $0.3. Pick kimi-k3 when you need a single response far longer than 65,536 tokens, since its max output stretches to 1,048,576, and when its reasoning capability flag matters.
Pricing
| Gemini 3.6 Flash | Kimi K3 | Δ | |
|---|---|---|---|
| Input / 1M tokens | $1.5 | $3 | 0.5× |
| Output / 1M tokens | $7.5 | $15 | 0.5× |
| Cache read / 1M tokens | $0.15 | $0.3 | 0.5× |
Rates from the live catalog at build time; each model page carries the current card.
Where they sit — input price per 1M tokens across all 63 chat models on this billing unit (log scale)
Capabilities
| Gemini 3.6 Flash | Kimi K3 | |
|---|---|---|
| Tool use | yes | yes |
| Thinking control | always on | always on |
| Structured output | yes | yes |
| Prompt caching | implicit + explicit | implicit (automatic) |
| Cache lifetime | not published | not published |
| Minimum cached prefix | 4096 tokens | not published |
Specs
| Gemini 3.6 Flash | Kimi K3 | |
|---|---|---|
| Input modalities | text image audio video | text image video |
| Output modalities | text | text |
| Knowledge cutoff | 2026-03 | — |
| Context window | 1M | 1M |
| Max output | 66K | 1M |
| Thinking parameter | thinkingLevel | reasoning_effort (top-level; the thinking object is not accepted) |
| Accepted values |
| reasoning_effort
|
| Default | medium | max |
Specs are transcribed from each vendor’s documentation; a row a vendor does not publish is left out rather than inferred. Full sources: Gemini 3.6 Flash · Kimi K3
One prompt, both models — measured through the gateway
Gemini 3.6 Flash passed · 3 sentences
Enabling reasoning causes the model to generate thousands of internal "thinking" tokens behind the scenes to process logic before delivering the final response. Because our vendor bills us for every token generated—including these intermediate processing steps—our billable output volume per query surged significantly. Since output tokens carry a much higher unit cost than input tokens, this jump in processing volume directly doubled our overall spend.
out 893 tok (+816 thinking) latency 4.7 s
Kimi K3 passed · 3 sentences
Reasoning models don't just answer questions—they "think" first, generating long internal chains of step-by-step logic before producing a response. Those hidden thinking steps are billed as output tokens (the most expensive kind, typically 3–5x the price of input tokens), and a single query can generate thousands of them even when the visible answer is only a paragraph long. So you're paying for dramatically more compute per request: the bill doubled because the model does far more work behind the scenes, not because usage increased.
out 755 tok (+637 thinking) latency 20.8 s
Instruction following (exactly three sentences — countable), audience shaping (CFO register), and the hidden-thinking billing gap the token meters below expose.
Gemini 3.6 Flash passed · 8/8 cases
Here is the fixed function along with an explanation of the bugs: ```python def longest_run(xs): if not xs: return 0 best = 1 cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 else: cur = 1 best = max(best, cur) return best ``` ### Explanation The original function initialized `cur` and `best` to `0` and incremented `cur` on matching adjacent items, which meant it was counting identical adjacent *pairs* rather than the total number of *elements* in a run. This caused single-element inputs to skip the loop and return `0` instead of `1`, and undercounted all run lengths by 1 (for example, a run of two items at the end of the list resulted in `cur = 1`). […]
out 2034 tok (+1768 thinking) latency 10.0 s
Kimi K3 passed · 8/8 cases
```python def longest_run(xs): if not xs: return 0 best = 1 cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 best = max(best, cur) else: cur = 1 return best ``` The bug is a fencepost error: `cur` counts *matching adjacent pairs* rather than *items in the run*, because it starts at 0 and only increments when `xs[i] == xs[i - 1]`. A run of length L contains L−1 equal pairs, so every run was undercounted by exactly one. For a single-element list (a run of length 1) the loop never runs and the function returns 0 instead of 1; likewise any run ending at the last index — e.g. the two `2`s in `[1, 2, 2]` — reported 1 instead of 2, making it look like no run existed at all. […]
out 1837 tok (+1547 thinking) latency 47.2 s
Whether the fix is actually correct (runnable), explanation density, and token efficiency on a bounded task.
Gemini 3.6 Flash passed · 5/5 fields, guidance "withheld"
```json { "company": "Meridian Robotics", "quarter": "September quarter", "revenue_usd_m": 412, "growth_pct": null, "guidance": "withheld" } ```
out 2843 tok (+2783 thinking) latency 13.1 s
Kimi K3 passed · 5/5 fields, guidance null
{ "company": "Meridian Robotics", "quarter": "September quarter", "revenue_usd_m": 412, "growth_pct": 11, "guidance": null }
out 924 tok (+863 thinking) latency 25.9 s
Schema obedience (no invented fields), hallucination pressure (guidance is explicitly withheld), and structured-output path differences.
Gemini 3.6 Flash missed · 60 words, 0 banned, 0 questions
72: reducing 73: your 74: monthly 75: token 76: spend. S6 (21): 77: You 78: can 79: easily 80: set 81: custom 82: expiration 83: rules, 84: configure 85: TTL 86: settings, 87: and 88: manage 89: cache 90: invalidation 91: across 92: all 93: vendors 94: from 95: one 96: centralized 97: dashboard. S7 (23): 98: Start 99: optimizing […]
out 4092 tok (+3929 thinking) latency 16.3 s
Kimi K3 passed · 120 words, 0 banned, 1 question
Meet Prompt Cache, a new API gateway capability that stores prompt responses and serves them across OpenAI, Anthropic, Google, and Azure endpoints. It matches requests by model, prompt hash, tools, temperature, and tenant policy, so repeated work returns fast while sensitive contexts stay isolated. Teams set TTLs, stale rules, encryption scopes, and bypass flags per route. Analytics show hit rate, latency saved, token spend avoided, and drift risk by provider. What changes for developers? Keep one integration, add cache headers, and watch fallback logic respect consent, residency, and audit needs. During rollout, canary keys compare fresh answers with cached copies before promotion. Prompt Cache cuts vendor calls, steadies p95 latency, and gives platform owners controls for cost, quality, and compliance.
out 1527 tok (+1354 thinking) latency 37.9 s
Constraint obedience (word budget, banned-word list, the single question), style fingerprint, and length control.
Switch between them with one line
Both ids are in every tab below — the highlighted pair of lines is the only edit. Same endpoint, same key, same request shape.
from openai import OpenAI
client = OpenAI(
base_url="https://synthorai.io/v1",
api_key="sk-syn-...",
)
resp = client.chat.completions.create(
model="gemini-3.6-flash",
# model="kimi-k3", # uncomment this line, comment the one above
messages=[{"role": "user", "content": "Summarize this diff"}],
)
print(resp.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://synthorai.io/v1",
apiKey: "sk-syn-...",
});
const resp = await client.chat.completions.create({
model: "gemini-3.6-flash",
// model: "kimi-k3", // uncomment this line, comment the one above
messages: [{ role: "user", content: "Summarize this diff" }],
});
console.log(resp.choices[0].message.content);curl https://synthorai.io/v1/chat/completions \
-H "Authorization: Bearer sk-syn-..." \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-3.6-flash",
# "model": "kimi-k3", # uncomment this line, comment the one above
"messages": [{"role": "user", "content": "Hello"}]
}'package main
import (
"context"
"fmt"
"github.com/openai/openai-go/v3"
"github.com/openai/openai-go/v3/option"
)
func main() {
client := openai.NewClient(
option.WithBaseURL("https://synthorai.io/v1"),
option.WithAPIKey("sk-syn-..."),
)
resp, _ := client.Chat.Completions.New(context.TODO(), openai.ChatCompletionNewParams{
Model: "gemini-3.6-flash",
// Model: "kimi-k3", // uncomment this line, comment the one above
Messages: []openai.ChatCompletionMessageParamUnion{
openai.UserMessage("Summarize this diff"),
},
})
fmt.Println(resp.Choices[0].Message.Content)
}import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.chat.completions.*;
OpenAIClient client = OpenAIOkHttpClient.builder()
.baseUrl("https://synthorai.io/v1")
.apiKey("sk-syn-...")
.build();
ChatCompletion resp = client.chat().completions().create(
ChatCompletionCreateParams.builder()
.model("gemini-3.6-flash")
// .model("kimi-k3") // uncomment this line, comment the one above
.addUserMessage("Summarize this diff")
.build());
System.out.println(resp.choices().get(0).message().content().orElse(""));FAQ
Which is cheaper, Gemini 3.6 Flash or Kimi K3?
Gemini 3.6 Flash is cheaper on input / 1m tokens ($1.5 vs $3, 2.0× apart). Other rows may point the other way — the table above carries the full card, and real cost depends on your mix.
Can I A/B test Gemini 3.6 Flash against Kimi K3 without two integrations?
Yes. Both are served through the same OpenAI-compatible endpoint with one API key — switching is a one-line model-string change, so you can route a fraction of traffic to each and compare bills directly.
Do Gemini 3.6 Flash and Kimi K3 support prompt caching?
Yes — both bill cached reads below their input rate, so warm-prefix workloads cost less than the list rates suggest. The exact cache-read rows are in the pricing table above.