SiliconFlow: Models Intelligence, Performance & Price
This analysis is intended to support you in choosing the best model provided by SiliconFlow for your use-case.
Most Intelligent
GLM-5.3-Flash
DeepSeek V4.1 Flash (max)DeepSeek V4 Pro 0813 (max)DeepSeek V4 Flash Vision (max)DeepSeek V4 Flash 0731 (max)Intelligence Index
Total 37 models
Fastest
DeepSeek V4 Flash Vision (max)DeepSeek V4.1 Flash (max)
Gemma 4 12B (non-reasoning)Gemma 4 12BGLM-5.2 (max) (FP8)Output speed
Total 37 models
Lowest Price
DeepSeek V4 Flash (high) (FP8)DeepSeek V4 Flash (max) (FP8)DeepSeek V4.1 Flash (max)GLM-5.3-Flash
Hy3 (FP8)Blended price (per 1M tokens)
Total 37 models
SiliconFlow offers 37 models, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across models.
- For intelligence, the top models on SiliconFlow are GLM-5.3-Flash (42), DeepSeek V4.1 Flash (max) (39), and DeepSeek V4 Pro 0813 (max) (36).
- For output speed, the fastest models are DeepSeek V4 Flash Vision (max) (223 t/s), DeepSeek V4.1 Flash (max) (161 t/s), and Gemma 4 12B (non-reasoning) (117 t/s). Speed varies significantly across models, with a 111% difference between the fastest and slowest.
- For latency, GLM-5.2 (non-reasoning) (FP8) (1.39s), Kimi K2.6 (non-reasoning) (FP8) (1.55s), and Gemma 4 26B A4B (non-reasoning) (FP8) (2.20s) offer the lowest time to first answer token.
- For pricing, DeepSeek V4 Flash (high) (FP8) ($0.07), DeepSeek V4 Flash (max) (FP8) ($0.07), and DeepSeek V4.1 Flash (max) ($0.09) offer the lowest blended prices per 1M tokens.
- For context window size, GLM-5.2 (max) (FP8) (1M), DeepSeek V4 Pro (max) (FP8) (1M), and DeepSeek V4 Pro (high) (FP8) (1M) support the largest context windows on SiliconFlow.
Intelligence Evaluations
Artificial Analysis Intelligence Index
Intelligence Evaluations
Agentic knowledge work, (Elo-500)/2000
Agentic real-world work tasks, (Elo-500)/2000
Agentic SaaS workflows
Agentic coding & terminal use
Coding
Reasoning & knowledge
Professional document reasoning, All-pass
Physics reasoning
Knowledge
1 - hallucination rate
Long context reasoning
Legal agentic work, Hallucination-Gated All-Pass Rate
Agentic business operations
Agentic scientific research workflows in a terminal
Quantitative analysis on spreadsheets & documents
Kubernetes incident root-cause analysis
Visual reasoning
Medical long context reasoning
Intelligence Index vs. Price
Context Window
Context Window
Pricing
Intelligence Index vs. Price
Performance Summary
Output Speed vs. Price
Speed
Measured by Output Speed (tokens per second)
Output Speed
Latency
Measured by Time (seconds) to First Token
Latency: Time To First Answer Token
Cache Behaviour
Cache Hit Rate
Cost per Task vs. Cache Hit Rate
End-to-End Response Time
Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed
End-to-End Response Time vs. Price
Further Analysis | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
GLM-5.3-Flash | 1M | Open | 42 | $0.50 | 55 | 1.55 | 47.04 | 36.39 | |||
DeepSeek V4.1 Flash (max) | 1M | Open | 39 | $0.86 | 159 | 2.07 | 17.83 | 12.60 | |||
DeepSeek V4 Pro 0813 (max) | 1.05M | Open | 36 | $1.45 | 84 | 1.58 | 31.48 | 23.93 | |||
DeepSeek V4 Flash Vision (max) | 1M | Proprietary | 35 | $2.51 | 216 | 1.26 | 12.84 | 9.26 | |||
DeepSeek V4 Flash 0731 (max) | 1.05M | Open | 34 | $0.15 | 43 | 2.29 | 60.27 | 46.38 | |||
GLM-5.2 (max) (FP8) | 1.05M | Open | 34 | $0.85 | 110 | 1.38 | 24.18 | 18.24 | |||
DeepSeek V4 Pro (max) (FP8) | 1.05M | Open | 30 | -- | 48 | 1.81 | 102.49 | 90.35 | |||
DeepSeek V4 Pro (high) (FP8) | 1.05M | Open | 30* | -- | 48 | 2.04 | 54.32 | 41.78 | |||
 MiniMax-M3 (FP8) | 1M | Open | 29 | $0.45 | 96 | 2.38 | 28.55 | 20.94 | |||
GLM-5 (FP8) | 200k | Open | 28* | -- | -- | -- | -- | -- | |||
 Kimi K2.6 (FP8) | 262k | Open | 27 | $0.61 | 26 | 1.57 | 192.23 | 171.40 | |||
GLM-5.1 (FP8) | 205k | Open | 26 | $1.55 | 30 | 2.31 | 147.12 | 127.93 | |||
Hy3 (FP8) | 256k | Open | 25 | $0.07 | 87 | 3.06 | 31.73 | 22.93 | |||
DeepSeek V4 Flash (high) (FP8) | 1.05M | Open | 24 | $0.12 | 42 | 1.52 | 43.08 | 29.62 | |||
GLM-5.1 (non-reasoning) (FP8) | 205k | Open | 24* | -- | 29 | 2.39 | 19.89 | -- | |||
DeepSeek V4 Flash (max) (FP8) | 1.05M | Open | 24 | $0.10 | 45 | 1.48 | 136.29 | 123.78 | |||
Kimi K2.6 (non-reasoning) (FP8) | 262k | Open | 24* | -- | 27 | 1.57 | 20.18 | -- | |||
Kimi K2.5 (FP8) | 262k | Open | 23* | -- | 27 | 2.21 | 132.89 | 111.84 | |||
 Qwen3.5 27B (FP8) | 262k | Open | 23* | -- | 22 | 3.48 | 114.72 | 88.99 | |||
GLM-5.2 (non-reasoning) (FP8) | 1.05M | Open | 22* | -- | 92 | 1.40 | 6.84 | -- | |||
GLM-5 (non-reasoning) (FP8) | 205k | Open | 22* | -- | -- | -- | -- | -- | |||
DeepSeek V3.2 (FP8) | 164k | Open | 21* | -- | 16 | 2.68 | 159.85 | 125.73 | |||
Qwen3.6 27B (FP8) | 262k | Open | 21 | $0.36 | 36 | 3.17 | 173.11 | 156.18 | |||
Qwen3.5 35B A3B (FP8) | 262k | Open | 19* | -- | 41 | 2.08 | 62.39 | 48.25 | |||
 LongCat 2.0 (FP8) | 262k | Open | 19 | $0.15 | 67 | 2.61 | 39.77 | 29.73 | |||
Qwen3.6 35B A3B (FP8) | 262k | Open | 18 | $0.32 | 81 | 1.85 | 74.50 | 66.48 | |||
 Step 3.5 Flash (FP8) | 262k | Open | 17* | -- | 76 | 1.59 | 34.49 | 26.32 | |||
DeepSeek V3.2 (non-reasoning) (FP8) | 164k | Open | 16* | -- | 15 | 3.02 | 35.64 | -- | |||
Qwen3.5 122B A10B (FP8) | 262k | Open | 16 | $0.21 | 56 | 1.77 | 46.81 | 36.03 | |||
Gemma 4 31B (FP8) | 262k | Open | 15 | $0.54 | 47 | 3.62 | 51.49 | 37.16 | |||
Gemma 4 12B | 262k | Open | 14* | -- | 114 | 2.42 | 24.38 | 17.57 | |||
Gemma 4 31B (non-reasoning) (FP8) | 262k | Open | 14* | -- | 48 | 3.50 | 14.00 | -- | |||
Gemma 4 26B A4B (non-reasoning) (FP8) | 262k | Open | 13* | -- | 87 | 2.29 | 8.05 | -- | |||
 Seed-OSS-36B-Instruct | 262k | Open | 12* | -- | 40 | 2.86 | 65.28 | 49.94 | |||
GLM-4.6V | 128k | Open | 11* | -- | -- | -- | -- | -- | |||
Qwen3.5 9B (FP8) | 262k | Open | 11 | $0.16 | 29 | 2.41 | 87.72 | 68.25 | |||
GLM-4.5-Air | 98.3k | Open | 11* | -- | 81 | 2.48 | 33.46 | 24.78 | |||
Gemma 4 12B (non-reasoning) | 262k | Open | 9* | -- | 114 | 2.39 | 6.79 | -- | |||
GLM-4.6V (non-reasoning) | 128k | Open | 8* | -- | -- | -- | -- | -- | |||
 Ling-flash-2.0 | 131k | Open | 8* | -- | 4 | 2.52 | 134.67 | -- | |||
Qwen2.5 72B (FP8) | 32k | Open | 8* | -- | 29 | 4.15 | 21.18 | -- | |||
 ERNIE 4.5 300B A47B | 131k | Open | 8* | -- | -- | -- | -- | -- | |||
Ring-flash-2.0 | 131k | Open | 7* | -- | -- | -- | -- | -- | |||
Key definitions
Frequently Asked Questions
Common questions about SiliconFlow
SiliconFlow offers 37 models that we track: GLM-5.3-Flash, DeepSeek V4.1 Flash (max), DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash Vision (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max) (FP8), DeepSeek V4 Pro (max) (FP8), DeepSeek V4 Pro (high) (FP8), MiniMax-M3 (FP8), Kimi K2.6 (FP8), GLM-5.1 (FP8), Hy3 (FP8), DeepSeek V4 Flash (high) (FP8), GLM-5.1 (non-reasoning) (FP8), DeepSeek V4 Flash (max) (FP8), Kimi K2.6 (non-reasoning) (FP8), Kimi K2.5 (FP8), Qwen3.5 27B (FP8), GLM-5.2 (non-reasoning) (FP8), DeepSeek V3.2 (FP8), Qwen3.6 27B (FP8), Qwen3.5 35B A3B (FP8), LongCat 2.0 (FP8), Qwen3.6 35B A3B (FP8), Step 3.5 Flash (FP8), DeepSeek V3.2 (non-reasoning) (FP8), Qwen3.5 122B A10B (FP8), Gemma 4 31B (FP8), Gemma 4 12B, Gemma 4 31B (non-reasoning) (FP8), Gemma 4 26B A4B (non-reasoning) (FP8), Seed-OSS-36B-Instruct, Qwen3.5 9B (FP8), GLM-4.5-Air, Gemma 4 12B (non-reasoning), Ling-flash-2.0, and Qwen2.5 72B (FP8).
The most intelligent model available on SiliconFlow is GLM-5.3-Flash with an Intelligence Index score of 42.
The fastest model on SiliconFlow by output speed is DeepSeek V4 Flash Vision (max) at 223.2 tokens per second.
The model with the lowest time to first answer token on SiliconFlow is GLM-5.2 (non-reasoning) (FP8) at 1.39s. Lower latency means faster initial response time.
The most affordable model on SiliconFlow by blended price is DeepSeek V4 Flash (high) (FP8) at $0.07 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on SiliconFlow vary up to 14x across models, from $0.07 per 1M tokens for DeepSeek V4 Flash (high) (FP8) to $1.03 per 1M tokens for GLM-5.1 (non-reasoning) (FP8).
Yes, SiliconFlow offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
30 of 37 models on SiliconFlow support JSON mode for structured output.
Yes, all 37 models on SiliconFlow support function calling (tool use).
Yes, SiliconFlow offers 28 reasoning models: GLM-5.3-Flash, DeepSeek V4.1 Flash (max), DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash Vision (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max) (FP8), DeepSeek V4 Pro (max) (FP8), DeepSeek V4 Pro (high) (FP8), MiniMax-M3 (FP8), Kimi K2.6 (FP8), GLM-5.1 (FP8), Hy3 (FP8), DeepSeek V4 Flash (high) (FP8), DeepSeek V4 Flash (max) (FP8), Kimi K2.5 (FP8), Qwen3.5 27B (FP8), DeepSeek V3.2 (FP8), Qwen3.6 27B (FP8), Qwen3.5 35B A3B (FP8), LongCat 2.0 (FP8), Qwen3.6 35B A3B (FP8), Step 3.5 Flash (FP8), Qwen3.5 122B A10B (FP8), Gemma 4 31B (FP8), Gemma 4 12B, Seed-OSS-36B-Instruct, Qwen3.5 9B (FP8), and GLM-4.5-Air. Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, 36 of 37 models on SiliconFlow are open weight models: GLM-5.3-Flash, DeepSeek V4.1 Flash (max), DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max) (FP8), DeepSeek V4 Pro (max) (FP8), DeepSeek V4 Pro (high) (FP8), MiniMax-M3 (FP8), Kimi K2.6 (FP8), GLM-5.1 (FP8), Hy3 (FP8), DeepSeek V4 Flash (high) (FP8), GLM-5.1 (non-reasoning) (FP8), DeepSeek V4 Flash (max) (FP8), Kimi K2.6 (non-reasoning) (FP8), Kimi K2.5 (FP8), Qwen3.5 27B (FP8), GLM-5.2 (non-reasoning) (FP8), DeepSeek V3.2 (FP8), Qwen3.6 27B (FP8), Qwen3.5 35B A3B (FP8), LongCat 2.0 (FP8), Qwen3.6 35B A3B (FP8), Step 3.5 Flash (FP8), DeepSeek V3.2 (non-reasoning) (FP8), Qwen3.5 122B A10B (FP8), Gemma 4 31B (FP8), Gemma 4 12B, Gemma 4 31B (non-reasoning) (FP8), Gemma 4 26B A4B (non-reasoning) (FP8), Seed-OSS-36B-Instruct, Qwen3.5 9B (FP8), GLM-4.5-Air, Gemma 4 12B (non-reasoning), Ling-flash-2.0, and Qwen2.5 72B (FP8).
Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.
When choosing a model on SiliconFlow, consider: intelligence (for quality-sensitive tasks), output speed (for throughput-intensive tasks), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and features like context window size, JSON mode, or function calling support.