DeepSeek V4 Flash (0731) vs Kimi K3
Which one, when — curated verdict, not a benchmark table
Both are text-out reasoning models with chat, code and tools at around a million tokens of context (1000000 for deepseek-v4-flash-0731, 1048576 for kimi-k3), so the split is cost and inputs: deepseek-v4-flash-0731 runs $0.308 input and $0.924 output, while kimi-k3 charges about 9.7x more on input ($3) and roughly 16x more on output ($15). Pick kimi-k3 when you need image or video input, or a single reply up to 1048576 tokens; pick deepseek-v4-flash-0731 for high-volume text work, where its 393216 max output and $0.0308 cache reads keep costs low.
Pricing
| DeepSeek V4 Flash (0731) | Kimi K3 | Δ | |
|---|---|---|---|
| Input / 1M tokens | $0.308 | $3 | 0.1× |
| Output / 1M tokens | $0.924 | $15 | 0.062× |
| Cache read / 1M tokens | $0.0308 | $0.3 | 0.1× |
| Cache write | no separate charge | — | — |
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
| DeepSeek V4 Flash (0731) | Kimi K3 | |
|---|---|---|
| Tool use | yes | yes |
| Thinking control | yes — vendor dial not published | always on |
| Structured output | yes | yes |
| Prompt caching | implicit (automatic) | implicit (automatic) |
| Cache lifetime | no fixed TTL (evicted when unused) | not published |
| Minimum cached prefix | not published | not published |
Specs
| DeepSeek V4 Flash (0731) | Kimi K3 | |
|---|---|---|
| Input modalities | text | text image video |
| Output modalities | text | text |
| Released | 2026-07-31 | — |
| Context window | 1M | 1M |
| Max output | 393K | 1M |
| Thinking parameter | — | reasoning_effort (top-level; the thinking object is not accepted) |
| Accepted values | — | reasoning_effort
|
| Default | — | max |
Specs are transcribed from each vendor’s documentation; a row a vendor does not publish is left out rather than inferred. Full sources: DeepSeek V4 Flash (0731) · Kimi K3
One prompt, both models — measured through the gateway
DeepSeek V4 Flash (0731) passed · 3 sentences
Enabling reasoning mode prompts the model to generate extensive internal chain-of-thought tokens before answering, often multiplying total token output per query by several times. This directly inflates costs because we are billed per token, and these reasoning tokens are processed at a premium compute tier that carries a higher price per million tokens. However, this spend buys measurably higher accuracy and fewer failed calls, so we are tracking it against specific error-reduction KPIs to ensure the ROI justifies the increase.
out 365 tok (+266 thinking) latency 5.1 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.
DeepSeek V4 Flash (0731) 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 original bug stems from initializing `cur` to 0 and only incrementing it when adjacent elements are equal, which counts the number of *equal transitions* rather than the actual run length—so every run of length L is reported as L−1, and a single‑element run (with no transitions) yields 0. Moreover, because `best` is updated only inside the equality branch, it never accounts for a run that ends at the last element if that run is longer than any previously seen (though the real issue remains the off‑by‑one). […]
out 1268 tok (+1016 thinking) latency 14.0 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.
DeepSeek V4 Flash (0731) passed · 5/5 fields, guidance null
{ "company": "Meridian Robotics", "quarter": "Q3", "revenue_usd_m": 412, "growth_pct": null, "guidance": null }
out 1119 tok (+1072 thinking) latency 12.3 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.
DeepSeek V4 Flash (0731) missed · 141 words, 0 banned, 1 question
**Introducing Prompt Cache for API Gateway** We are pleased to announce Prompt Cache, a new feature that stores and reuses identical prompt prefixes across all connected LLM providers. By automatically recognizing repeated requests, this cache reduces token consumption and lowers latency without altering your application logic. Whether you use OpenAI, Anthropic, or open-source models, the cache operates transparently, ensuring consistent behavior while cutting costs. Developers can configure cache TTL and scoping rules via a simple API call, giving you precise control over data freshness and privacy. This feature also minimizes redundant upstream calls, freeing provider rate limits for more critical workloads. Start caching today through the dashboard or CLI, and watch your operational expenses drop significantly. […]
out 254 tok (+80 thinking) latency 4.4 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="deepseek-v4-flash-0731",
# model="kimi-k3", # uncomment this line, comment the one above
messages=[{"role": "user", "content": "Summarize this diff"}],
reasoning_effort="medium",
)
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: "deepseek-v4-flash-0731",
// model: "kimi-k3", // uncomment this line, comment the one above
messages: [{ role: "user", content: "Summarize this diff" }],
reasoning_effort: "medium",
});
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": "deepseek-v4-flash-0731",
# "model": "kimi-k3", # uncomment this line, comment the one above
"messages": [{"role": "user", "content": "Hello"}],
"reasoning_effort": "medium"
}'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: "deepseek-v4-flash-0731",
// Model: "kimi-k3", // uncomment this line, comment the one above
Messages: []openai.ChatCompletionMessageParamUnion{
openai.UserMessage("Summarize this diff"),
},
ReasoningEffort: openai.ReasoningEffortMedium,
})
fmt.Println(resp.Choices[0].Message.Content)
}import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.chat.completions.*;
import com.openai.models.ReasoningEffort;
OpenAIClient client = OpenAIOkHttpClient.builder()
.baseUrl("https://synthorai.io/v1")
.apiKey("sk-syn-...")
.build();
ChatCompletion resp = client.chat().completions().create(
ChatCompletionCreateParams.builder()
.model("deepseek-v4-flash-0731")
// .model("kimi-k3") // uncomment this line, comment the one above
.addUserMessage("Summarize this diff")
.reasoningEffort(ReasoningEffort.MEDIUM)
.build());
System.out.println(resp.choices().get(0).message().content().orElse(""));FAQ
Which is cheaper, DeepSeek V4 Flash (0731) or Kimi K3?
DeepSeek V4 Flash (0731) is cheaper on input / 1m tokens ($0.308 vs $3, 9.7× 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 DeepSeek V4 Flash (0731) 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 DeepSeek V4 Flash (0731) 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.