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
DeepSeek V4.1 Flash 1017 (max)
GLM-5.3 (max)
MiMo-V2.6-ProGLM-5.3-FlashMiMo-V2.6-FlashIntelligence Index
Total 106 models
Fastest
Nemotron 3.5 Super VL
Inkling SmallInkling (xhigh) (FP8)Nemotron 3 NanoNVIDIA Nemotron Nano 9B V2 (non-reasoning)Output speed
Total 106 models
Lowest Price
Llama 3.1 8B (Turbo, FP8)Llama 3.1 8B
Gemma 4 E4B (non-reasoning)Gemma 4 E4BMiMo-V2.6-FlashBlended price (per 1M tokens)
Total 106 models
DeepInfra offers 106 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 DeepSeek V4.1 Flash 1017 (max) (32), GLM-5.3 (max) (32), and MiMo-V2.6-Pro (30).
- For output speed, the fastest models are Nemotron 3.5 Super VL (580 t/s), Inkling Small (322 t/s), and Inkling (xhigh) (FP8) (309 t/s). Speed varies significantly across models, with a 144% difference between the fastest and slowest.
- For latency, NVIDIA Nemotron Nano 9B V2 (non-reasoning) (0.51s), Nemotron 3 Nano (non-reasoning) (0.54s), and Gemma 4 26B A4B (non-reasoning) (FP8) (0.56s) 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 Gemma 4 E4B (non-reasoning) ($0.03) offer the lowest blended prices per 1M tokens.
- For context window size, Llama 4 Maverick (FP8) (1M), DeepSeek V4 Pro (max) (FP4) (1M), and DeepSeek V4 Flash (max) (FP4) (1M) support the largest context windows on DeepInfra.
Intelligence Evaluations
Artificial Analysis Intelligence Index
Evaluations
Agentic knowledge work, Elo
Knowledge
1 - hallucination rate
Agentic SaaS workflows
Agentic scientific research workflows in a terminal
Agentic software engineering in open-source repositories
Reasoning & knowledge
Professional document reasoning, All-pass
Agentic coding & terminal use
Physics reasoning
Quantitative analysis on spreadsheets & documents
Legal agentic work, Hallucination-Gated All-Pass Rate
Agentic business operations
Agentic real-world work tasks, Elo
Agentic interactive 3D experiences, Elo
Product and brand videos designed in code, Elo
Kubernetes incident root-cause analysis
Medical long context reasoning
Visual reasoning
Rich, varied, well-crafted language
Well-structured, scannable, answer-first
Truthful, acknowledges uncertainty, non-sycophantic
Use of filler and generic language
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 | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
DeepSeek V4.1 Flash 1017 (max) | 1M | Open | 32 | $0.19 | 66 | 0.77 | 38.64 | 30.29 | |||
GLM-5.3 (max) | 1.05M | Open | 32 | $2.39 | 78 | 1.28 | 33.42 | 25.71 | |||
MiMo-V2.6-Pro | 1M | Open | 30 | $0.55 | 54 | 1.15 | 47.29 | 36.91 | |||
GLM-5.3-Flash | 1M | Open | 28 | $1.16 | 99 | 0.93 | 26.11 | 20.14 | |||
MiMo-V2.6-Flash | 1M | Open | 27 | $0.08 | 74 | 1.56 | 35.20 | 26.92 | |||
DeepSeek V4 Pro 0813 (max) | 1M | Open | 23 | $2.98 | 67 | 1.00 | 38.06 | 29.64 | |||
 Qwen3.8 27B (xhigh) | 262k | Open | 20 | $1.54 | 38 | 1.16 | 66.89 | 52.59 | |||
 MiniMax-M3 | 524k | Open | 19 | $1.11 | 23 | 0.77 | 111.10 | 88.27 | |||
Nemotron 3.5 Super VL | 1M | Open | 14 | $0.35 | 580 | 0.81 | 5.12 | 3.45 | |||
Inkling Small | 524k | Open | 13 | $0.29 | 322 | 0.58 | 8.35 | 6.22 | |||
Inkling (xhigh) (FP8) | 131k | Open | 12 | $1.26 | 309 | 0.60 | 8.69 | 6.47 | |||
Nemotron 3 Ultra | 262k | Open | 12 | $1.74 | 48 | 1.44 | 59.01 | 47.20 | |||
Nemotron 3 Ultra BF16 | 262k | Open | 12 | $4.03 | 67 | 1.38 | 43.08 | 34.19 | |||
Muse Glimmer (high) | 131k | Open | 6 | $0.13 | 102 | 0.89 | 25.48 | 19.68 | |||
Nemotron 3.5 Lightning (NVFP4) | 262k | Open | 6 | $0.31 | 231 | 0.53 | 11.37 | 8.67 | |||
MiMo-V2.5-Pro | 65.5k | Open | -- | -- | 56 | 1.06 | 45.43 | 35.49 | |||
Qwen3.5 122B A10B (non-reasoning) (FP4) | 262k | Open | -- | -- | 59 | 1.09 | 9.59 | -- | |||
Qwen3 Coder 480B (Turbo, FP4) | 262k | Open | -- | -- | 53 | 0.98 | 10.42 | -- | |||
 Kimi K2.6 (FP4) | 262k | Open | -- | -- | 104 | 1.01 | 48.77 | 42.94 | |||
Qwen3.5 397B A17B (FP8) | 262k | Open | -- | -- | 58 | 1.15 | 64.40 | 54.67 | |||
Kimi K2.6 (non-reasoning) (FP4) | 262k | Open | -- | -- | 73 | 0.91 | 7.76 | -- | |||
MiniMax-M2.5 (FP8) | 197k | Open | -- | -- | 23 | 0.96 | 109.82 | 87.09 | |||
Qwen3.5 27B (FP8) | 262k | Open | -- | -- | 68 | 1.08 | 37.67 | 29.27 | |||
 Hermes 3 - Llama-3.1 70B | 131k | Open | -- | -- | 29 | 2.52 | 19.77 | -- | |||
Gemma 3 4B | 131k | Open | -- | -- | 74 | 0.80 | 7.53 | -- | |||
Qwen3.5 122B A10B (FP4) | 262k | Open | -- | -- | 53 | 1.09 | 48.37 | 37.82 | |||
MiMo-V2.5 | 262k | Open | -- | -- | 59 | 1.29 | 43.54 | 33.80 | |||
Nemotron 3 Nano (non-reasoning) | 262k | Open | -- | -- | 206 | 0.54 | 2.96 | -- | |||
 Mistral Small 3.1 | 128k | Open | -- | -- | 44 | 1.02 | 12.42 | -- | |||
Qwen3 Next 80B A3B | 262k | Open | -- | -- | 106 | 0.89 | 5.60 | -- | |||
Gemma 3 27B | 131k | Open | -- | -- | 27 | 1.48 | 19.81 | -- | |||
GLM-4.7-Flash | 203k | Open | -- | -- | 76 | 1.07 | 33.80 | 26.18 | |||
Gemma 3 12B | 131k | Open | -- | -- | 28 | 1.13 | 19.11 | -- | |||
NVIDIA Nemotron Nano 9B V2 (non-reasoning) | 131k | Open | -- | -- | 238 | 0.51 | 2.61 | -- | |||
Qwen3.5 4B (non-reasoning) FP8 | 262k | Open | -- | -- | 15 | 0.99 | 34.41 | -- | |||
Qwen3.5 397B A17B (non-reasoning) (FP8) | 262k | Open | -- | -- | 57 | 1.14 | 9.90 | -- | |||
Qwen3 235B 2507 (FP8) | 262k | Open | -- | -- | 13 | 0.87 | 39.04 | -- | |||
 gpt-oss-20b (high) | 131k | Open | -- | -- | 130 | 0.53 | 19.81 | 15.43 | |||
Qwen3.6 27B (non-reasoning) FP8 | 262k | Open | -- | -- | 61 | 1.09 | 9.26 | -- | |||
Qwen3 30B (FP8) | 41k | Open | -- | -- | 118 | 0.58 | 21.75 | 16.93 | |||
GLM-5 (FP4) | 203k | Open | -- | -- | 78 | 1.08 | 47.40 | 39.90 | |||
Mistral Small 3.2 (FP8) | 128k | Open | -- | -- | 51 | 1.02 | 10.83 | -- | |||
DeepSeek V3 (Dec) | 164k | Open | -- | -- | 19 | 0.97 | 27.46 | -- | |||
Qwen3 14B (FP8) | 32.8k | Open | -- | -- | 87 | 0.88 | 29.70 | 23.06 | |||
Llama 3.1 70B | 131k | Open | -- | -- | 39 | 2.32 | 15.01 | -- | |||
Llama 3.1 70B (Turbo, FP8) | 131k | Open | -- | -- | 41 | 2.04 | 14.15 | -- | |||
Qwen3.5 35B A3B (FP8) | 262k | Open | -- | -- | 49 | 0.63 | 51.62 | 40.79 | |||
MiMo-V2.5-Pro (non-reasoning) | 65.5k | Open | -- | -- | 50 | 1.30 | 11.29 | -- | |||
DeepSeek R1 Distill Llama 70B | 131k | Open | -- | -- | -- | -- | -- | -- | |||
Qwen3.5 27B (non-reasoning) FP8 | 262k | Open | -- | -- | 69 | 1.18 | 8.39 | -- | |||
Gemma 4 E4B (non-reasoning) | 131k | Open | -- | -- | 44 | 0.87 | 12.18 | -- | |||
GLM-5.1 (FP4) | 203k | Open | -- | -- | 80 | 1.11 | 54.68 | 47.33 | |||
DeepSeek V4 Pro (high) (FP4) | 65.5k | Open | -- | -- | -- | -- | -- | -- | |||
Llama 3.2 11B (Vision) | 131k | Open | -- | -- | 30 | 1.28 | 17.76 | -- | |||
Nemotron 3 Super | 262k | Open | -- | -- | 81 | 8.80 | 39.67 | 24.69 | |||
Qwen3.5 4B (FP8) | 262k | Open | -- | -- | 13 | 0.87 | 187.59 | 149.38 | |||
Qwen3.6 35B A3B (non-reasoning) (FP8) | 262k | Open | -- | -- | 88 | 0.79 | 6.45 | -- | |||
GLM-4.6 (FP4) | 203k | Open | -- | -- | 22 | 1.33 | 116.06 | 91.79 | |||
 Step 3.7 Flash | 256k | Open | -- | -- | 63 | 2.32 | 41.73 | 31.52 | |||
DeepSeek V3.2 (non-reasoning) | 164k | Open | -- | -- | 32 | 1.04 | 16.71 | -- | |||
Gemma 4 26B A4B (FP8) | 262k | Open | -- | -- | 78 | 0.71 | 32.79 | 25.66 | |||
GLM-4.7 (FP4) | 203k | Open | -- | -- | 29 | 1.39 | 86.50 | 68.08 | |||
Gemma 4 26B A4B (non-reasoning) (FP8) | 262k | Open | -- | -- | 73 | 0.56 | 7.37 | -- | |||
Llama 3.1 8B (Turbo, FP8) | 131k | Open | -- | -- | 32 | 0.92 | 16.45 | -- | |||
Llama 3.1 8B | 131k | Open | -- | -- | 34 | 0.96 | 15.83 | -- | |||
DeepSeek V3 0324 (FP4) | 164k | Open | -- | -- | 139 | 0.72 | 4.32 | -- | |||
Nemotron 3 Nano | 262k | Open | -- | -- | 240 | 0.52 | 10.94 | 8.33 | |||
DeepSeek V4 Flash Vision (max) | 1M | Proprietary | -- | -- | 83 | 0.84 | 30.86 | 24.02 | |||
GLM-4.7 (non-reasoning) (FP4) | 203k | Open | -- | -- | 28 | 1.26 | 18.93 | -- | |||
Llama 3 8B | 8.19k | Open | -- | -- | -- | -- | -- | -- | |||
Qwen3.6 35B A3B (FP8) | 262k | Open | -- | -- | 37 | 0.87 | 160.26 | 145.87 | |||
DeepSeek V4 Flash (high) (FP4) | 1.05M | Open | -- | -- | -- | -- | -- | -- | |||
Qwen3.6 27B FP8 | 262k | Open | -- | -- | 58 | 1.17 | 107.91 | 98.10 | |||
Kimi K2.7 Code | 262k | Open | -- | -- | 83 | 1.11 | 34.09 | 26.94 | |||
Qwen3.8 2.4T A95B | 262k | Open | -- | -- | 86 | 0.87 | 29.83 | 23.17 | |||
Llama 4 Maverick (FP8) | 1.05M | Open | -- | -- | 44 | 0.83 | 12.12 | -- | |||
GLM-5.1 (non-reasoning) (FP4) | 203k | Open | -- | -- | 94 | 0.97 | 6.30 | -- | |||
Llama 3.3 70B (Turbo, FP8) | 131k | Open | -- | -- | 19 | 2.14 | 28.96 | -- | |||
Kimi K2.5 | 262k | Open | -- | -- | 95 | 0.88 | 37.43 | 31.28 | |||
 Ling 3.1 Flash | 262k | Open | -- | -- | 26 | 0.94 | 97.31 | 77.10 | |||
DeepSeek R1 0528 | 164k | Open | -- | -- | 30 | 1.13 | 85.64 | 67.61 | |||
Llama 4 Scout | 131k | Open | -- | -- | 19 | 0.90 | 26.77 | -- | |||
 Granite 4.2 8B | 131k | Open | -- | -- | 57 | 1.05 | 44.64 | 34.87 | |||
Qwen3 32B (FP8) | 41k | Open | -- | -- | 26 | 1.40 | 99.28 | 78.31 | |||
 Hy3 (FP8) | 262k | Open | -- | -- | 45 | 1.93 | 57.64 | 44.57 | |||
Qwen2.5 72B | 32.8k | Open | -- | -- | 25 | 2.22 | 22.23 | -- | |||
DeepSeek V4 Pro (max) (FP4) | 1.05M | Open | -- | -- | 56 | 1.02 | 88.48 | 78.49 | |||
Qwen3 32B (non-reasoning) (FP8) | 41k | Open | -- | -- | 27 | 1.40 | 20.15 | -- | |||
gpt-oss-120b (low) | 131k | Open | -- | -- | 39 | 0.70 | 65.07 | 51.49 | |||
Gemma 4 31B | 262k | Open | -- | -- | 26 | 1.73 | 89.32 | 68.01 | |||
Gemma 4 31B (non-reasoning) (FP8) | 262k | Open | -- | -- | 23 | 1.49 | 23.57 | -- | |||
DeepSeek V3.2 (FP4) | 164k | Open | -- | -- | 28 | 1.02 | 90.62 | 71.68 | |||
Gemma 4 E4B | 131k | Open | -- | -- | 43 | 0.98 | 59.26 | 46.62 | |||
DeepSeek V4 Flash (max) (FP4) | 1.05M | Open | -- | -- | 31 | 1.52 | 199.88 | 182.13 | |||
DeepSeek V4.1 Flash (max) | 1M | Open | -- | -- | 121 | 0.72 | 21.43 | 16.57 | |||
DeepSeek V3.1 Terminus (non-reasoning) (FP4) | 164k | Open | -- | -- | 57 | 0.81 | 9.52 | -- | |||
Ling 3.0 Flash | 131k | Open | -- | -- | 48 | 0.76 | 52.31 | 41.25 | |||
Mistral Small 3 | 32.8k | Open | -- | -- | 50 | 1.04 | 11.05 | -- | |||
Qwen3 14B (non-reasoning) (FP8) | 41k | Open | -- | -- | 87 | 0.85 | 6.58 | -- | |||
 Phi-4 | 16.4k | Open | -- | -- | 104 | 1.08 | 5.88 | -- | |||
Qwen3.5 35B A3B (non-reasoning) FP8 | 262k | Open | -- | -- | 67 | 0.64 | 8.14 | -- | |||
gpt-oss-120b (high) (Turbo) | 131k | Open | -- | -- | 36 | 0.75 | 69.46 | 54.97 | |||
gpt-oss-120b (high) | 131k | Open | -- | -- | 39 | 0.77 | 65.16 | 51.52 | |||
GLM-5 (non-reasoning) (FP8) | 203k | Open | -- | -- | 68 | 1.00 | 8.36 | -- | |||
NVIDIA Nemotron Nano 9B V2 | 131k | Open | -- | -- | 213 | 0.53 | 12.28 | 9.40 | |||
Qwen3 30B (non-reasoning) (FP8) | 41k | Open | -- | -- | 123 | 0.57 | 4.65 | -- | |||
Qwen3 235B A22B 2507 (FP8) | 262k | Open | -- | -- | 44 | 0.87 | 57.42 | 45.24 | |||
GLM-5.2 (max) (FP4) | 1.05M | Open | -- | -- | 71 | 1.05 | 36.44 | 28.31 | |||
DeepSeek V3.1 (non-reasoning) (FP4) | 164k | Open | -- | -- | 32 | 1.12 | 16.69 | -- | |||
DeepSeek V4 Flash 0731 (max) | 1M | Open | -- | -- | 54 | 0.74 | 47.41 | 37.33 | |||
Key definitions
Frequently Asked Questions
Common questions about DeepInfra
DeepInfra offers 106 models that we track: DeepSeek V4.1 Flash 1017 (max), GLM-5.3 (max), MiMo-V2.6-Pro, GLM-5.3-Flash, MiMo-V2.6-Flash, DeepSeek V4 Pro 0813 (max), Qwen3.8 27B (xhigh), MiniMax-M3, Nemotron 3.5 Super VL, Inkling Small, Inkling (xhigh) (FP8), Nemotron 3 Ultra, Nemotron 3 Ultra BF16, Muse Glimmer (high), Nemotron 3.5 Lightning (NVFP4), MiMo-V2.5-Pro, Qwen3.5 122B A10B (non-reasoning) (FP4), Qwen3 Coder 480B (Turbo, FP4), Kimi K2.6 (FP4), Qwen3.5 397B A17B (FP8), Kimi K2.6 (non-reasoning) (FP4), MiniMax-M2.5 (FP8), Qwen3.5 27B (FP8), Hermes 3 - Llama-3.1 70B, Gemma 3 4B, Qwen3.5 122B A10B (FP4), MiMo-V2.5, Nemotron 3 Nano (non-reasoning), Mistral Small 3.1, Qwen3 Next 80B A3B, Gemma 3 27B, GLM-4.7-Flash, Gemma 3 12B, NVIDIA Nemotron Nano 9B V2 (non-reasoning), Qwen3.5 4B (non-reasoning) FP8, Qwen3.5 397B A17B (non-reasoning) (FP8), Qwen3 235B 2507 (FP8), gpt-oss-20b (high), Qwen3.6 27B (non-reasoning) FP8, Qwen3 30B (FP8), GLM-5 (FP4), Mistral Small 3.2 (FP8), DeepSeek V3 (Dec), Qwen3 14B (FP8), Llama 3.1 70B, Llama 3.1 70B (Turbo, FP8), Qwen3.5 35B A3B (FP8), MiMo-V2.5-Pro (non-reasoning), Qwen3.5 27B (non-reasoning) FP8, Gemma 4 E4B (non-reasoning), GLM-5.1 (FP4), Llama 3.2 11B (Vision), Nemotron 3 Super, Qwen3.5 4B (FP8), Qwen3.6 35B A3B (non-reasoning) (FP8), GLM-4.6 (FP4), Step 3.7 Flash, DeepSeek V3.2 (non-reasoning), Gemma 4 26B A4B (FP8), GLM-4.7 (FP4), Gemma 4 26B A4B (non-reasoning) (FP8), Llama 3.1 8B (Turbo, FP8), Llama 3.1 8B, DeepSeek V3 0324 (FP4), Nemotron 3 Nano, DeepSeek V4 Flash Vision (max), GLM-4.7 (non-reasoning) (FP4), Qwen3.6 35B A3B (FP8), Qwen3.6 27B FP8, Kimi K2.7 Code, Qwen3.8 2.4T A95B, Llama 4 Maverick (FP8), GLM-5.1 (non-reasoning) (FP4), Llama 3.3 70B (Turbo, FP8), Kimi K2.5, Ling 3.1 Flash, DeepSeek R1 0528, Llama 4 Scout, Granite 4.2 8B, Qwen3 32B (FP8), Hy3 (FP8), Qwen2.5 72B, DeepSeek V4 Pro (max) (FP4), Qwen3 32B (non-reasoning) (FP8), gpt-oss-120b (low), Gemma 4 31B, Gemma 4 31B (non-reasoning) (FP8), DeepSeek V3.2 (FP4), Gemma 4 E4B, DeepSeek V4 Flash (max) (FP4), DeepSeek V4.1 Flash (max), DeepSeek V3.1 Terminus (non-reasoning) (FP4), Ling 3.0 Flash, Mistral Small 3, Qwen3 14B (non-reasoning) (FP8), Phi-4, Qwen3.5 35B A3B (non-reasoning) FP8, gpt-oss-120b (high) (Turbo), gpt-oss-120b (high), GLM-5 (non-reasoning) (FP8), NVIDIA Nemotron Nano 9B V2, Qwen3 30B (non-reasoning) (FP8), Qwen3 235B A22B 2507 (FP8), GLM-5.2 (max) (FP4), DeepSeek V3.1 (non-reasoning) (FP4), and DeepSeek V4 Flash 0731 (max).
The most intelligent model available on DeepInfra is DeepSeek V4.1 Flash 1017 (max) with an Intelligence Index score of 32.
The fastest model on DeepInfra by output speed is Nemotron 3.5 Super VL at 579.6 tokens per second.
The model with the lowest time to first answer token on DeepInfra is NVIDIA Nemotron Nano 9B V2 (non-reasoning) at 0.51s. 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 106 models on DeepInfra support JSON mode for structured output.
103 of 106 models on DeepInfra support function calling (tool use).
Yes, DeepInfra offers 61 reasoning models: DeepSeek V4.1 Flash 1017 (max), GLM-5.3 (max), MiMo-V2.6-Pro, GLM-5.3-Flash, MiMo-V2.6-Flash, DeepSeek V4 Pro 0813 (max), Qwen3.8 27B (xhigh), MiniMax-M3, Nemotron 3.5 Super VL, Inkling Small, Inkling (xhigh) (FP8), Nemotron 3 Ultra, Nemotron 3 Ultra BF16, Muse Glimmer (high), Nemotron 3.5 Lightning (NVFP4), MiMo-V2.5-Pro, Kimi K2.6 (FP4), Qwen3.5 397B A17B (FP8), MiniMax-M2.5 (FP8), Qwen3.5 27B (FP8), Qwen3.5 122B A10B (FP4), MiMo-V2.5, GLM-4.7-Flash, gpt-oss-20b (high), Qwen3 30B (FP8), GLM-5 (FP4), Qwen3 14B (FP8), Qwen3.5 35B A3B (FP8), GLM-5.1 (FP4), Nemotron 3 Super, Qwen3.5 4B (FP8), GLM-4.6 (FP4), Step 3.7 Flash, Gemma 4 26B A4B (FP8), GLM-4.7 (FP4), Nemotron 3 Nano, DeepSeek V4 Flash Vision (max), Qwen3.6 35B A3B (FP8), Qwen3.6 27B FP8, Kimi K2.7 Code, Qwen3.8 2.4T A95B, Kimi K2.5, Ling 3.1 Flash, DeepSeek R1 0528, Granite 4.2 8B, Qwen3 32B (FP8), Hy3 (FP8), DeepSeek V4 Pro (max) (FP4), gpt-oss-120b (low), Gemma 4 31B, DeepSeek V3.2 (FP4), Gemma 4 E4B, DeepSeek V4 Flash (max) (FP4), DeepSeek V4.1 Flash (max), Ling 3.0 Flash, gpt-oss-120b (high) (Turbo), gpt-oss-120b (high), NVIDIA Nemotron Nano 9B V2, Qwen3 235B A22B 2507 (FP8), GLM-5.2 (max) (FP4), and DeepSeek V4 Flash 0731 (max). Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, 105 of 106 models on DeepInfra are open weight models: DeepSeek V4.1 Flash 1017 (max), GLM-5.3 (max), MiMo-V2.6-Pro, GLM-5.3-Flash, MiMo-V2.6-Flash, DeepSeek V4 Pro 0813 (max), Qwen3.8 27B (xhigh), MiniMax-M3, Nemotron 3.5 Super VL, Inkling Small, Inkling (xhigh) (FP8), Nemotron 3 Ultra, Nemotron 3 Ultra BF16, Muse Glimmer (high), Nemotron 3.5 Lightning (NVFP4), MiMo-V2.5-Pro, Qwen3.5 122B A10B (non-reasoning) (FP4), Qwen3 Coder 480B (Turbo, FP4), Kimi K2.6 (FP4), Qwen3.5 397B A17B (FP8), Kimi K2.6 (non-reasoning) (FP4), MiniMax-M2.5 (FP8), Qwen3.5 27B (FP8), Hermes 3 - Llama-3.1 70B, Gemma 3 4B, Qwen3.5 122B A10B (FP4), MiMo-V2.5, Nemotron 3 Nano (non-reasoning), Mistral Small 3.1, Qwen3 Next 80B A3B, Gemma 3 27B, GLM-4.7-Flash, Gemma 3 12B, NVIDIA Nemotron Nano 9B V2 (non-reasoning), Qwen3.5 4B (non-reasoning) FP8, Qwen3.5 397B A17B (non-reasoning) (FP8), Qwen3 235B 2507 (FP8), gpt-oss-20b (high), Qwen3.6 27B (non-reasoning) FP8, Qwen3 30B (FP8), GLM-5 (FP4), Mistral Small 3.2 (FP8), DeepSeek V3 (Dec), Qwen3 14B (FP8), Llama 3.1 70B, Llama 3.1 70B (Turbo, FP8), Qwen3.5 35B A3B (FP8), MiMo-V2.5-Pro (non-reasoning), Qwen3.5 27B (non-reasoning) FP8, Gemma 4 E4B (non-reasoning), GLM-5.1 (FP4), Llama 3.2 11B (Vision), Nemotron 3 Super, Qwen3.5 4B (FP8), Qwen3.6 35B A3B (non-reasoning) (FP8), GLM-4.6 (FP4), Step 3.7 Flash, DeepSeek V3.2 (non-reasoning), Gemma 4 26B A4B (FP8), GLM-4.7 (FP4), Gemma 4 26B A4B (non-reasoning) (FP8), Llama 3.1 8B (Turbo, FP8), Llama 3.1 8B, DeepSeek V3 0324 (FP4), Nemotron 3 Nano, GLM-4.7 (non-reasoning) (FP4), Qwen3.6 35B A3B (FP8), Qwen3.6 27B FP8, Kimi K2.7 Code, Qwen3.8 2.4T A95B, Llama 4 Maverick (FP8), GLM-5.1 (non-reasoning) (FP4), Llama 3.3 70B (Turbo, FP8), Kimi K2.5, Ling 3.1 Flash, DeepSeek R1 0528, Llama 4 Scout, Granite 4.2 8B, Qwen3 32B (FP8), Hy3 (FP8), Qwen2.5 72B, DeepSeek V4 Pro (max) (FP4), Qwen3 32B (non-reasoning) (FP8), gpt-oss-120b (low), Gemma 4 31B, Gemma 4 31B (non-reasoning) (FP8), DeepSeek V3.2 (FP4), Gemma 4 E4B, DeepSeek V4 Flash (max) (FP4), DeepSeek V4.1 Flash (max), DeepSeek V3.1 Terminus (non-reasoning) (FP4), Ling 3.0 Flash, Mistral Small 3, Qwen3 14B (non-reasoning) (FP8), Phi-4, Qwen3.5 35B A3B (non-reasoning) FP8, gpt-oss-120b (high) (Turbo), gpt-oss-120b (high), GLM-5 (non-reasoning) (FP8), NVIDIA Nemotron Nano 9B V2, Qwen3 30B (non-reasoning) (FP8), Qwen3 235B A22B 2507 (FP8), GLM-5.2 (max) (FP4), DeepSeek V3.1 (non-reasoning) (FP4), and DeepSeek V4 Flash 0731 (max).
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.