DeepInfra: Models Intelligence, Performance & Price
This analysis is intended to support you in choosing the best model provided by DeepInfra for your use-case.
Most Intelligent
MiMo-V2.6-Pro
GLM-5.3 (max)GLM-5.3-Flash
Qwen3.8 2.4T A95B
DeepSeek V4.1 Flash (max)Intelligence Index
Total 107 models
Fastest
Nemotron 3.5 Lightning (NVFP4)
Inkling SmallNVIDIA Nemotron Nano 9B V2
Granite 4.2 3BNemotron 3 NanoOutput speed
Total 107 models
Lowest Price
Llama 3.1 8B (Turbo, FP8)Llama 3.1 8BGranite 4.2 3B
Gemma 4 E4BGemma 4 E4B (non-reasoning)Blended price (per 1M tokens)
Total 107 models
DeepInfra offers 107 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 DeepInfra are MiMo-V2.6-Pro (46), GLM-5.3 (max) (45), and GLM-5.3-Flash (42).
- For output speed, the fastest models are Nemotron 3.5 Lightning (NVFP4) (263 t/s), Inkling Small (231 t/s), and NVIDIA Nemotron Nano 9B V2 (222 t/s).
- For latency, Nemotron 3 Nano (non-reasoning) (0.53s), NVIDIA Nemotron Nano 9B V2 (non-reasoning) (0.56s), and Qwen3.5 35B A3B (non-reasoning) FP8 (0.58s) offer the lowest time to first answer token.
- For pricing, Llama 3.1 8B (Turbo, FP8) ($0.02), Llama 3.1 8B ($0.02), and Granite 4.2 3B ($0.02) offer the lowest blended prices per 1M tokens.
- For context window size, GLM-5.2 (max) (FP4) (1M), DeepSeek V4 Pro (max) (FP4) (1M), and DeepSeek V4 Flash (high) (FP4) (1M) support the largest context windows on DeepInfra.
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 | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
MiMo-V2.6-Pro | 1M | Open | 46 | $0.96 | 33 | 1.27 | 77.82 | 61.24 | |||
GLM-5.3 (max) | 1.05M | Open | 45 | $0.97 | 78 | 1.34 | 33.40 | 25.65 | |||
GLM-5.3-Flash | 1M | Open | 42 | $0.34 | 27 | 1.43 | 92.69 | 73.01 | |||
Qwen3.8 2.4T A95B | 262k | Open | 40 | $2.89 | 112 | 1.22 | 23.61 | 17.91 | |||
DeepSeek V4.1 Flash (max) | 1M | Open | 39 | $0.44 | 70 | 0.80 | 36.45 | 28.52 | |||
MiMo-V2.6-Flash | 1M | Open | 38 | $0.42 | 31 | 1.42 | 81.24 | 63.86 | |||
DeepSeek V4 Pro 0813 (max) | 1M | Open | 36 | $1.00 | 42 | 1.44 | 61.30 | 47.89 | |||
DeepSeek V4 Flash Vision (max) | 1M | Proprietary | 35 | $0.51 | 134 | 0.98 | 19.66 | 14.95 | |||
DeepSeek V4 Flash 0731 (max) | 1M | Open | 34 | $0.09 | 42 | 1.00 | 59.90 | 47.11 | |||
GLM-5.2 (max) (FP4) | 1.05M | Open | 34 | $0.37 | 62 | 1.26 | 41.31 | 32.04 | |||
Qwen3.8 27B (xhigh) | 262k | Open | 34 | $0.52 | 53 | 0.92 | 47.70 | 37.42 | |||
DeepSeek V4 Pro (max) (FP4) | 1.05M | Open | 30 | $1.19 | 65 | 1.32 | 75.97 | 67.00 | |||
DeepSeek V4 Pro (high) (FP4) | 65.5k | Open | 30* | -- | 51 | 1.28 | 49.70 | 38.71 | |||
 MiniMax-M3 | 524k | Open | 29 | $0.44 | 23 | 1.16 | 107.79 | 85.30 | |||
GLM-5 (FP4) | 203k | Open | 28* | -- | 63 | 1.36 | 59.02 | 49.66 | |||
 Kimi K2.6 (FP4) | 262k | Open | 27 | $0.51 | 18 | 1.42 | 279.27 | 249.79 | |||
GLM-5.1 (FP4) | 203k | Open | 26 | $0.58 | 71 | 1.17 | 61.23 | 53.07 | |||
MiMo-V2.5-Pro | 65.5k | Open | 26 | $0.28 | 35 | 1.53 | 72.64 | 56.89 | |||
Kimi K2.7 Code | 262k | Open | 26 | $0.39 | 53 | 1.26 | 53.00 | 42.27 | |||
Inkling Small | 524k | Open | 26 | -- | 135 | 0.85 | 19.31 | 14.77 | |||
 Hy3 (FP8) | 262k | Open | 25 | $0.07 | 55 | 1.03 | 46.21 | 36.14 | |||
MiMo-V2.5 | 262k | Open | 25* | -- | 36 | 1.25 | 69.88 | 54.91 | |||
Inkling (xhigh) (FP8) | 131k | Open | 25 | -- | 110 | 0.74 | 23.45 | 18.17 | |||
DeepSeek V4 Flash (high) (FP4) | 1.05M | Open | 24 | $0.08 | 28 | 1.27 | 64.34 | 44.94 | |||
GLM-5.1 (non-reasoning) (FP4) | 203k | Open | 24* | -- | 89 | 1.12 | 6.74 | -- | |||
DeepSeek V4 Flash (max) (FP4) | 1.05M | Open | 24 | $0.07 | 28 | 1.75 | 222.39 | 202.59 | |||
Kimi K2.6 (non-reasoning) (FP4) | 262k | Open | 24* | -- | 15 | 1.59 | 35.18 | -- | |||
Kimi K2.5 | 262k | Open | 23* | -- | 19 | 1.46 | 185.51 | 157.51 | |||
Nemotron 3 Ultra | 262k | Open | 23 | $0.39 | 32 | 1.71 | 89.36 | 71.86 | |||
Nemotron 3 Ultra BF16 | 262k | Open | 23 | $0.99 | 42 | 1.64 | 67.80 | 54.24 | |||
Qwen3.5 27B (FP8) | 262k | Open | 23* | -- | 56 | 1.22 | 45.90 | 35.74 | |||
MiniMax-M2.5 (FP8) | 197k | Open | 23* | -- | 22 | 1.23 | 115.08 | 91.08 | |||
GLM-4.7 (FP4) | 203k | Open | 22* | -- | 18 | 1.16 | 137.40 | 109.00 | |||
GLM-5 (non-reasoning) (FP8) | 203k | Open | 22* | -- | 70 | 1.12 | 8.31 | -- | |||
DeepSeek V3.2 (FP4) | 164k | Open | 21* | -- | 20 | 1.55 | 126.17 | 99.69 | |||
Qwen3.5 397B A17B (non-reasoning) (FP8) | 262k | Open | 21* | -- | 28 | 1.30 | 18.86 | -- | |||
Qwen3.6 27B FP8 | 262k | Open | 21 | $0.37 | 52 | 1.03 | 120.86 | 110.13 | |||
 Ling 3.0 Flash | 131k | Open | 20 | $0.02 | 44 | 0.64 | 57.71 | 45.65 | |||
Qwen3.6 27B (non-reasoning) FP8 | 262k | Open | 20* | -- | 49 | 1.14 | 11.33 | -- | |||
 Step 3.7 Flash | 256k | Open | 19* | -- | 39 | 1.47 | 65.59 | 51.30 | |||
Qwen3.5 27B (non-reasoning) FP8 | 262k | Open | 19* | -- | 49 | 1.24 | 11.51 | -- | |||
Qwen3.5 35B A3B (FP8) | 262k | Open | 19* | -- | 65 | 0.60 | 38.87 | 30.62 | |||
GLM-4.6 (FP4) | 203k | Open | 19* | -- | 23 | 1.58 | 110.75 | 87.34 | |||
Qwen3.5 397B A17B (FP8) | 262k | Open | 18 | -- | 35 | 1.27 | 107.73 | 92.02 | |||
MiMo-V2.5-Pro (non-reasoning) | 65.5k | Open | 18* | -- | 30 | 1.32 | 17.88 | -- | |||
Qwen3.6 35B A3B (FP8) | 262k | Open | 18 | $0.14 | 38 | 0.94 | 158.12 | 143.85 | |||
Qwen3.5 122B A10B (non-reasoning) (FP4) | 262k | Open | 18* | -- | 30 | 1.49 | 18.30 | -- | |||
Muse Glimmer (high) | 131k | Open | 17 | $0.05 | 75 | 1.41 | 34.55 | 26.51 | |||
GLM-4.7 (non-reasoning) (FP4) | 203k | Open | 17* | -- | 19 | 1.44 | 27.45 | -- | |||
Gemma 4 26B A4B (FP8) | 262k | Open | 17* | -- | 42 | 0.84 | 60.62 | 47.82 | |||
DeepSeek V3.2 (non-reasoning) | 164k | Open | 16* | -- | 20 | 1.51 | 26.22 | -- | |||
Qwen3.5 122B A10B (FP4) | 262k | Open | 16 | $0.24 | 28 | 1.44 | 91.94 | 72.40 | |||
Qwen3.6 35B A3B (non-reasoning) (FP8) | 262k | Open | 15* | -- | 80 | 0.81 | 7.08 | -- | |||
Qwen3.5 35B A3B (non-reasoning) FP8 | 262k | Open | 15* | -- | 75 | 0.60 | 7.24 | -- | |||
GLM-4.7-Flash | 203k | Open | 15* | -- | 26 | 1.52 | 97.97 | 77.16 | |||
Gemma 4 31B | 262k | Open | 15 | $0.15 | 20 | 2.57 | 112.18 | 85.10 | |||
DeepSeek V3.1 Terminus (non-reasoning) (FP4) | 164k | Open | 14* | -- | 39 | 0.92 | 13.84 | -- | |||
Gemma 4 31B (non-reasoning) (FP8) | 262k | Open | 14* | -- | 18 | 2.26 | 29.91 | -- | |||
DeepSeek V3.1 (non-reasoning) (FP4) | 164k | Open | 14* | -- | 8 | 2.65 | 66.15 | -- | |||
Gemma 4 26B A4B (non-reasoning) (FP8) | 262k | Open | 13* | -- | 44 | 0.69 | 12.13 | -- | |||
Qwen3.5 4B (FP8) | 262k | Open | 13* | -- | 23 | 0.86 | 110.73 | 87.89 | |||
DeepSeek R1 0528 | 164k | Open | 13* | -- | 26 | 0.86 | 95.39 | 75.63 | |||
Nemotron 3.5 Lightning (NVFP4) | 262k | Open | 13 | $0.09 | 307 | 0.54 | 8.68 | 6.51 | |||
Nemotron 3 Super | 262k | Open | 13 | $0.48 | 54 | 15.04 | 61.65 | 37.29 | |||
Qwen3 235B A22B 2507 (FP8) | 262k | Open | 13 | -- | 89 | 0.80 | 29.03 | 22.58 | |||
Granite 4.2 30B | 131k | Open | 13 | $0.07 | 76 | 0.96 | 33.92 | 26.37 | |||
Qwen3 235B 2507 (FP8) | 262k | Open | 12* | -- | 11 | 1.47 | 46.22 | -- | |||
Qwen3 Coder 480B (Turbo, FP4) | 262k | Open | 12* | -- | 64 | 1.52 | 9.34 | -- | |||
 gpt-oss-120b (high) (Turbo) | 131k | Open | 12 | $0.11 | 191 | 0.89 | 13.97 | 10.46 | |||
gpt-oss-120b (high) | 131k | Open | 12 | $0.03 | 43 | 0.68 | 59.38 | 46.96 | |||
Granite 4.2 8B | 131k | Open | 11 | $0.02 | 52 | 0.75 | 48.92 | 38.54 | |||
Qwen3.5 4B (non-reasoning) FP8 | 262k | Open | 11* | -- | 25 | 0.84 | 20.58 | -- | |||
gpt-oss-120b (low) | 131k | Open | 10* | -- | 42 | 0.69 | 59.58 | 47.11 | |||
Llama 4 Maverick (FP8) | 1.05M | Open | 10* | -- | 25 | 0.84 | 20.68 | -- | |||
DeepSeek V3 0324 (FP4) | 164k | Open | 10 | $0.02 | 79 | 0.84 | 7.21 | -- | |||
Qwen3 Next 80B A3B | 262k | Open | 10* | -- | 130 | 0.80 | 4.66 | -- | |||
Granite 4.2 3B | 131k | Open | 9 | $0.01 | 215 | 0.53 | 12.14 | 9.29 | |||
gpt-oss-20b (high) | 131k | Open | 9 | $0.01 | 111 | 0.52 | 23.09 | 18.05 | |||
Gemma 4 E4B | 131k | Open | 9* | -- | 28 | 1.09 | 90.91 | 71.86 | |||
Nemotron 3 Nano | 262k | Open | 9 | $0.02 | 19 | 66.66 | 198.96 | 105.84 | |||
Qwen3 32B (FP8) | 41k | Open | 9* | -- | 29 | 1.57 | 88.73 | 69.72 | |||
DeepSeek V3 (Dec) | 164k | Open | 8 | $0.02 | 20 | 3.36 | 28.25 | -- | |||
 Mistral Small 3.2 (FP8) | 128k | Open | 8* | -- | 38 | 0.88 | 14.12 | -- | |||
Qwen3 14B (FP8) | 32.8k | Open | 8* | -- | 20 | 1.71 | 125.96 | 99.40 | |||
Llama 4 Scout | 131k | Open | 8* | -- | 26 | 0.79 | 19.79 | -- | |||
DeepSeek R1 Distill Llama 70B | 131k | Open | 8* | -- | -- | -- | -- | -- | |||
Qwen2.5 72B | 32.8k | Open | 8* | -- | 23 | 2.35 | 23.87 | -- | |||
Llama 3.3 70B (Turbo, FP8) | 131k | Open | 8* | -- | 17 | 2.30 | 31.81 | -- | |||
Qwen3 30B (FP8) | 41k | Open | 8* | -- | 128 | 0.57 | 20.11 | 15.63 | |||
Gemma 4 E4B (non-reasoning) | 131k | Open | 7* | -- | 28 | 0.78 | 18.63 | -- | |||
NVIDIA Nemotron Nano 9B V2 | 131k | Open | 7* | -- | 18 | 71.66 | 209.85 | 110.56 | |||
Qwen3 32B (non-reasoning) (FP8) | 41k | Open | 7* | -- | 29 | 1.45 | 18.88 | -- | |||
Mistral Small 3.1 | 128k | Open | 7 | $0.02 | 37 | 0.94 | 14.56 | -- | |||
Llama 3.1 8B (Turbo, FP8) | 131k | Open | 7* | -- | 33 | 0.98 | 16.28 | -- | |||
Llama 3.1 8B | 131k | Open | 7* | -- | 34 | 1.13 | 16.04 | -- | |||
Nemotron 3 Nano (non-reasoning) | 262k | Open | 7* | -- | 21 | 19.31 | 42.96 | -- | |||
NVIDIA Nemotron Nano 9B V2 (non-reasoning) | 131k | Open | 7* | -- | 21 | 57.88 | 82.22 | -- | |||
Qwen3 14B (non-reasoning) (FP8) | 41k | Open | 7* | -- | 30 | 1.76 | 18.66 | -- | |||
Mistral Small 3 | 32.8k | Open | 7* | -- | 51 | 0.87 | 10.59 | -- | |||
Qwen3 30B (non-reasoning) (FP8) | 41k | Open | 7* | -- | 115 | 0.52 | 4.86 | -- | |||
Llama 3.1 70B | 131k | Open | 7* | -- | 39 | 1.82 | 14.49 | -- | |||
Llama 3.1 70B (Turbo, FP8) | 131k | Open | 7* | -- | 42 | 1.89 | 13.85 | -- | |||
 Hermes 3 - Llama-3.1 70B | 131k | Open | 6* | -- | 31 | 2.12 | 18.38 | -- | |||
 Phi-4 | 16.4k | Open | 6* | -- | 71 | 0.96 | 7.98 | -- | |||
Llama 3.2 11B (Vision) | 131k | Open | 5* | -- | 15 | 1.96 | 35.01 | -- | |||
Gemma 3 27B | 131k | Open | 5 | $0.16 | 16 | 1.89 | 32.89 | -- | |||
Gemma 3 4B | 131k | Open | 5* | -- | 25 | 1.11 | 21.00 | -- | |||
Llama 3 8B | 8.19k | Open | 5* | -- | -- | -- | -- | -- | |||
Gemma 3 12B | 131k | Open | 4 | $0.13 | 62 | 1.09 | 9.22 | -- | |||
Key definitions
Frequently Asked Questions
Common questions about DeepInfra
DeepInfra offers 107 models that we track: MiMo-V2.6-Pro, GLM-5.3 (max), GLM-5.3-Flash, Qwen3.8 2.4T A95B, DeepSeek V4.1 Flash (max), MiMo-V2.6-Flash, DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash Vision (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max) (FP4), Qwen3.8 27B (xhigh), DeepSeek V4 Pro (max) (FP4), DeepSeek V4 Pro (high) (FP4), MiniMax-M3, GLM-5 (FP4), Kimi K2.6 (FP4), GLM-5.1 (FP4), MiMo-V2.5-Pro, Kimi K2.7 Code, Inkling Small, Hy3 (FP8), MiMo-V2.5, Inkling (xhigh) (FP8), DeepSeek V4 Flash (high) (FP4), GLM-5.1 (non-reasoning) (FP4), DeepSeek V4 Flash (max) (FP4), Kimi K2.6 (non-reasoning) (FP4), Kimi K2.5, Nemotron 3 Ultra, Nemotron 3 Ultra BF16, Qwen3.5 27B (FP8), MiniMax-M2.5 (FP8), GLM-4.7 (FP4), GLM-5 (non-reasoning) (FP8), DeepSeek V3.2 (FP4), Qwen3.5 397B A17B (non-reasoning) (FP8), Qwen3.6 27B FP8, Ling 3.0 Flash, Qwen3.6 27B (non-reasoning) FP8, Step 3.7 Flash, Qwen3.5 27B (non-reasoning) FP8, Qwen3.5 35B A3B (FP8), GLM-4.6 (FP4), Qwen3.5 397B A17B (FP8), MiMo-V2.5-Pro (non-reasoning), Qwen3.6 35B A3B (FP8), Qwen3.5 122B A10B (non-reasoning) (FP4), Muse Glimmer (high), GLM-4.7 (non-reasoning) (FP4), Gemma 4 26B A4B (FP8), DeepSeek V3.2 (non-reasoning), Qwen3.5 122B A10B (FP4), Qwen3.6 35B A3B (non-reasoning) (FP8), Qwen3.5 35B A3B (non-reasoning) FP8, GLM-4.7-Flash, Gemma 4 31B, DeepSeek V3.1 Terminus (non-reasoning) (FP4), Gemma 4 31B (non-reasoning) (FP8), DeepSeek V3.1 (non-reasoning) (FP4), Gemma 4 26B A4B (non-reasoning) (FP8), Qwen3.5 4B (FP8), DeepSeek R1 0528, Nemotron 3.5 Lightning (NVFP4), Nemotron 3 Super, Qwen3 235B A22B 2507 (FP8), Granite 4.2 30B, Qwen3 235B 2507 (FP8), Qwen3 Coder 480B (Turbo, FP4), gpt-oss-120b (high) (Turbo), gpt-oss-120b (high), Granite 4.2 8B, Qwen3.5 4B (non-reasoning) FP8, gpt-oss-120b (low), Llama 4 Maverick (FP8), DeepSeek V3 0324 (FP4), Qwen3 Next 80B A3B, Granite 4.2 3B, gpt-oss-20b (high), Gemma 4 E4B, Nemotron 3 Nano, Qwen3 32B (FP8), DeepSeek V3 (Dec), Mistral Small 3.2 (FP8), Qwen3 14B (FP8), Llama 4 Scout, Qwen2.5 72B, Llama 3.3 70B (Turbo, FP8), Qwen3 30B (FP8), Gemma 4 E4B (non-reasoning), NVIDIA Nemotron Nano 9B V2, Qwen3 32B (non-reasoning) (FP8), Mistral Small 3.1, Llama 3.1 8B (Turbo, FP8), Llama 3.1 8B, Nemotron 3 Nano (non-reasoning), NVIDIA Nemotron Nano 9B V2 (non-reasoning), Qwen3 14B (non-reasoning) (FP8), Mistral Small 3, Qwen3 30B (non-reasoning) (FP8), Llama 3.1 70B, Llama 3.1 70B (Turbo, FP8), Hermes 3 - Llama-3.1 70B, Phi-4, Llama 3.2 11B (Vision), Gemma 3 27B, Gemma 3 4B, and Gemma 3 12B.
The most intelligent model available on DeepInfra is MiMo-V2.6-Pro with an Intelligence Index score of 46.
The fastest model on DeepInfra by output speed is Nemotron 3.5 Lightning (NVFP4) at 263.5 tokens per second.
The model with the lowest time to first answer token on DeepInfra is Nemotron 3 Nano (non-reasoning) at 0.53s. Lower latency means faster initial response time.
The most affordable model on DeepInfra by blended price is Llama 3.1 8B (Turbo, FP8) at $0.02 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on DeepInfra vary up to 52x across models, from $0.02 per 1M tokens for Llama 3.1 8B (Turbo, FP8) to $1.14 per 1M tokens for Qwen3.8 2.4T A95B.
Yes, DeepInfra offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
Yes, all 107 models on DeepInfra support JSON mode for structured output.
104 of 107 models on DeepInfra support function calling (tool use).
Yes, DeepInfra offers 62 reasoning models: MiMo-V2.6-Pro, GLM-5.3 (max), GLM-5.3-Flash, Qwen3.8 2.4T A95B, DeepSeek V4.1 Flash (max), MiMo-V2.6-Flash, DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash Vision (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max) (FP4), Qwen3.8 27B (xhigh), DeepSeek V4 Pro (max) (FP4), DeepSeek V4 Pro (high) (FP4), MiniMax-M3, GLM-5 (FP4), Kimi K2.6 (FP4), GLM-5.1 (FP4), MiMo-V2.5-Pro, Kimi K2.7 Code, Inkling Small, Hy3 (FP8), MiMo-V2.5, Inkling (xhigh) (FP8), DeepSeek V4 Flash (high) (FP4), DeepSeek V4 Flash (max) (FP4), Kimi K2.5, Nemotron 3 Ultra, Nemotron 3 Ultra BF16, Qwen3.5 27B (FP8), MiniMax-M2.5 (FP8), GLM-4.7 (FP4), DeepSeek V3.2 (FP4), Qwen3.6 27B FP8, Ling 3.0 Flash, Step 3.7 Flash, Qwen3.5 35B A3B (FP8), GLM-4.6 (FP4), Qwen3.5 397B A17B (FP8), Qwen3.6 35B A3B (FP8), Muse Glimmer (high), Gemma 4 26B A4B (FP8), Qwen3.5 122B A10B (FP4), GLM-4.7-Flash, Gemma 4 31B, Qwen3.5 4B (FP8), DeepSeek R1 0528, Nemotron 3.5 Lightning (NVFP4), Nemotron 3 Super, Qwen3 235B A22B 2507 (FP8), Granite 4.2 30B, gpt-oss-120b (high) (Turbo), gpt-oss-120b (high), Granite 4.2 8B, gpt-oss-120b (low), Granite 4.2 3B, gpt-oss-20b (high), Gemma 4 E4B, Nemotron 3 Nano, Qwen3 32B (FP8), Qwen3 14B (FP8), Qwen3 30B (FP8), and NVIDIA Nemotron Nano 9B V2. Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, 106 of 107 models on DeepInfra are open weight models: MiMo-V2.6-Pro, GLM-5.3 (max), GLM-5.3-Flash, Qwen3.8 2.4T A95B, DeepSeek V4.1 Flash (max), MiMo-V2.6-Flash, DeepSeek V4 Pro 0813 (max), DeepSeek V4 Flash 0731 (max), GLM-5.2 (max) (FP4), Qwen3.8 27B (xhigh), DeepSeek V4 Pro (max) (FP4), DeepSeek V4 Pro (high) (FP4), MiniMax-M3, GLM-5 (FP4), Kimi K2.6 (FP4), GLM-5.1 (FP4), MiMo-V2.5-Pro, Kimi K2.7 Code, Inkling Small, Hy3 (FP8), MiMo-V2.5, Inkling (xhigh) (FP8), DeepSeek V4 Flash (high) (FP4), GLM-5.1 (non-reasoning) (FP4), DeepSeek V4 Flash (max) (FP4), Kimi K2.6 (non-reasoning) (FP4), Kimi K2.5, Nemotron 3 Ultra, Nemotron 3 Ultra BF16, Qwen3.5 27B (FP8), MiniMax-M2.5 (FP8), GLM-4.7 (FP4), GLM-5 (non-reasoning) (FP8), DeepSeek V3.2 (FP4), Qwen3.5 397B A17B (non-reasoning) (FP8), Qwen3.6 27B FP8, Ling 3.0 Flash, Qwen3.6 27B (non-reasoning) FP8, Step 3.7 Flash, Qwen3.5 27B (non-reasoning) FP8, Qwen3.5 35B A3B (FP8), GLM-4.6 (FP4), Qwen3.5 397B A17B (FP8), MiMo-V2.5-Pro (non-reasoning), Qwen3.6 35B A3B (FP8), Qwen3.5 122B A10B (non-reasoning) (FP4), Muse Glimmer (high), GLM-4.7 (non-reasoning) (FP4), Gemma 4 26B A4B (FP8), DeepSeek V3.2 (non-reasoning), Qwen3.5 122B A10B (FP4), Qwen3.6 35B A3B (non-reasoning) (FP8), Qwen3.5 35B A3B (non-reasoning) FP8, GLM-4.7-Flash, Gemma 4 31B, DeepSeek V3.1 Terminus (non-reasoning) (FP4), Gemma 4 31B (non-reasoning) (FP8), DeepSeek V3.1 (non-reasoning) (FP4), Gemma 4 26B A4B (non-reasoning) (FP8), Qwen3.5 4B (FP8), DeepSeek R1 0528, Nemotron 3.5 Lightning (NVFP4), Nemotron 3 Super, Qwen3 235B A22B 2507 (FP8), Granite 4.2 30B, Qwen3 235B 2507 (FP8), Qwen3 Coder 480B (Turbo, FP4), gpt-oss-120b (high) (Turbo), gpt-oss-120b (high), Granite 4.2 8B, Qwen3.5 4B (non-reasoning) FP8, gpt-oss-120b (low), Llama 4 Maverick (FP8), DeepSeek V3 0324 (FP4), Qwen3 Next 80B A3B, Granite 4.2 3B, gpt-oss-20b (high), Gemma 4 E4B, Nemotron 3 Nano, Qwen3 32B (FP8), DeepSeek V3 (Dec), Mistral Small 3.2 (FP8), Qwen3 14B (FP8), Llama 4 Scout, Qwen2.5 72B, Llama 3.3 70B (Turbo, FP8), Qwen3 30B (FP8), Gemma 4 E4B (non-reasoning), NVIDIA Nemotron Nano 9B V2, Qwen3 32B (non-reasoning) (FP8), Mistral Small 3.1, Llama 3.1 8B (Turbo, FP8), Llama 3.1 8B, Nemotron 3 Nano (non-reasoning), NVIDIA Nemotron Nano 9B V2 (non-reasoning), Qwen3 14B (non-reasoning) (FP8), Mistral Small 3, Qwen3 30B (non-reasoning) (FP8), Llama 3.1 70B, Llama 3.1 70B (Turbo, FP8), Hermes 3 - Llama-3.1 70B, Phi-4, Llama 3.2 11B (Vision), Gemma 3 27B, Gemma 3 4B, and Gemma 3 12B.
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 DeepInfra, 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.