kimi-k3
Checking current pricing…
Chat comparison
kimi-k3vsmistral-large-3:675b
Compare the things that matter before you build: current price, capabilities, and latest aggregated LLM7 statistics.
Current pricing
See the cost difference clearly
mistral-large-3:675b
Checking current pricing…
What's different
Only capabilities that differ between these models are listed. Green highlights the broader supported option or larger context window.
Context window
kimi-k3Better fit
1,000,000 tokens
mistral-large-3:675b
256,000 tokens
Input formats
kimi-k3
text
mistral-large-3:675b
text, image
Vision
kimi-k3
Not supported
mistral-large-3:675bBetter fit
Supported
JSON mode
kimi-k3Better fit
Supported
mistral-large-3:675b
Not supported
Reasoning
kimi-k3
Not supported
mistral-large-3:675bBetter fit
Supported
Capability scores
AI-estimated ratings on 30 criteria from head-to-head comparisons against 86 other text models in the LLM7 catalog, judged by TypeSafe Jev from public information (2026-10-20). Each score is the win rate against the field, relative to the strongest model for that criterion (100%). Not a benchmark.
Below 50%50–79%80–94%95%+
kimi-k3 · leads in 5
40%
Weak
mistral-large-3:675b · leads in 24
49%
Weak
Knowledge & reasoning
General knowledge
kimi-k340%mistral-large-3:675bLeads52%Logical reasoning
kimi-k322%mistral-large-3:675bLeads57%Mathematics
kimi-k339%mistral-large-3:675bLeads62%Scientific and technical depth
kimi-k332%mistral-large-3:675bLeads62%Planning and task decomposition
kimi-k323%mistral-large-3:675bLeads55%
Coding & tools
Code generation
kimi-k331%mistral-large-3:675bLeads49%Debugging and code review
kimi-k329%mistral-large-3:675bLeads56%Agentic coding
kimi-k333%mistral-large-3:675bLeads54%Tool and function calling
kimi-k335%mistral-large-3:675bLeads54%SQL and query writing
kimi-k335%mistral-large-3:675bLeads55%
Writing & language
Creative writing
kimi-k334%mistral-large-3:675bLeads50%Professional writing
kimi-k337%mistral-large-3:675bLeads49%Conversational quality
kimi-k334%mistral-large-3:675bLeads50%Multilingual ability
kimi-k361%mistral-large-3:675bLeads86%Translation
kimi-k346%mistral-large-3:675bLeads59%Summarization
kimi-k3Leads62%mistral-large-3:675b45%
Data & professional work
Data analysis
kimi-k343%mistral-large-3:675bLeads54%Professional domains
kimi-k334%mistral-large-3:675bLeads56%Information extraction
kimi-k3Leads51%mistral-large-3:675b33%Classification and labeling
kimi-k341%mistral-large-3:675bLeads54%Long-context handling
kimi-k3Leads85%mistral-large-3:675b44%
Reliability & control
Instruction following
kimi-k334%mistral-large-3:675bLeads36%Structured output
kimi-k3Leads50%mistral-large-3:675b9%Factual accuracy
kimi-k340%mistral-large-3:675bLeads52%Safety and alignment
kimi-k334%mistral-large-3:675bLeads45%Steerability
kimi-k329%mistral-large-3:675bLeads46%Output consistency
kimi-k348%mistral-large-3:675bLeads54%
Speed, cost & vision
Response speed
kimi-k3Leads35%mistral-large-3:675b34%Cost efficiency
kimi-k339%mistral-large-3:675bLeads48%Image understanding
kimi-k3No image inputmistral-large-3:675b17%
Latest LLM7 statistics
Recent aggregated percentages and response-time metrics from LLM7, not a global model benchmark or traffic disclosure.
Recent stability
kimi-k3
97.87%
mistral-large-3:675b
100%
Average response time
kimi-k3
12,694.11 ms
mistral-large-3:675b
17,950.15 ms
P95 API latency
kimi-k3
60,000 ms
mistral-large-3:675b
60,000 ms
Side-by-side trends
Shared charts make the observed differences easier to read. Metrics without enough history stay hidden.
Request success rate
Both models on the same scale
kimi-k3: 100%mistral-large-3:675b: 100%
Average API latency
Both models on the same scale
kimi-k3: 69,350 msmistral-large-3:675b: 15,337 ms
P95 API latency
Both models on the same scale
kimi-k3: 60,000 msmistral-large-3:675b: 30,000 ms
Estimate your workload
Published rates are loaded from the pricing API; per-request quotes are billed from actual provider usage.
Checking current pricing…
Checking current pricing…
Quick take
- kimi-k3 has a larger context window.
- mistral-large-3:675b supports image input while the other model does not.
- mistral-large-3:675b recorded higher recent LLM7 stability during this period.