Parasail: Models Intelligence, Performance & Price
This analysis is intended to support you in choosing the best model provided by Parasail for your use-case.
Most Intelligent
Intelligence Index
Total 31 models
Lowest Time per Task
Time per Intelligence Index task (minutes)
Total 31 models
Lowest Price
Blended price (per 1M tokens)
Total 31 models
Parasail offers 31 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 Parasail are Kimi K3.1 Pro (46), DeepSeek V4.1 Flash 1017 (max) (33), and GLM-5.3 (max) (32).
- For time per task, gpt-oss-120b (high) (6.8 min), GLM-5.3-Flash (7.2 min), and GLM-5.3 (max) (9.5 min) finish an Intelligence Index task fastest on Parasail.
- For latency, Qwen3 Next 80B A3B (0.99s), Qwen3 Coder Next (FP8) (1.04s), and Qwen3.6 35B A3B (non-reasoning) (FP8) (1.13s) offer the lowest time to first answer token.
- For pricing, Gemma 3 4B (FP8) ($0.05), Gemma 3 27B ($0.09), and MiniMax-M3.1-Flash ($0.10) offer the lowest blended prices per 1M tokens. Prices vary up to 2.2x across models.
- For context window size, Kimi K3.1 Pro (1M), Llama 4 Maverick (FP8) (1M), and DeepSeek V4 Flash (max) (FP8) (1M) support the largest context windows on Parasail.
Intelligence
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
Pricing
Cost per Intelligence Index Task
Intelligence Index vs. Price
Speed
Average of the single-request and multi-turn agentic workloads measured
Time per Intelligence Index Task
Quality
Endpoint Accuracy Index
Latency
Average of the single-request and multi-turn agentic workloads measured
Time to First Answer Token
Reliability
Share of requests the endpoint failed, across all the single-request and multi-turn agentic workloads measured
Endpoint Error Rate
Caching
Cache Hit Rate
Cost per Task vs. Cache Hit Rate
Further Analysis | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Kimi K3.1 Pro | 46 | $4.22 | 13.3 | 64.80 | 0.8% | 1M | Open | |||
DeepSeek V4.1 Flash 1017 (max) | 33 | $0.52 | 7.6 | 29.69 | 0.0% | 1M | Open | |||
GLM-5.3 (max) | 32 | $4.92 | 12.6 | 41.56 | 3.1% | 1M | Open | |||
MiniMax-M3.1-Flash | 29 | $0.44 | 14.2 | 29.19 | 3.0% | 1M | Open | |||
GLM-5.3-Flash | 28 | $0.64 | 8.8 | 19.49 | 10.2% | 1M | Open | |||
Qwen3.8 27B (xhigh) (FP8) | 20 | $1.78 | 23.2 | 70.73 | 2.3% | 262k | Open | |||
MiniMax-M3 (MXFP8) | 19 | $1.17 | 14.5 | 8.67 | 9.6% | 1M | Open | |||
Gemma 4 31B | 4 | -- | -- | 19.58 | 3.6% | 262k | Open | |||
gpt-oss-120b (high) | 4 | $0.17 | 7.1 | 15.92 | 1.7% | 131k | Open | |||
Kimi K2.6 | -- | -- | -- | 93.80 | 41.2% | 262k | Open | |||
Qwen3.5 397B A17B | -- | -- | -- | 36.30 | 3.6% | 262k | Open | |||
Kimi K2.6 (non-reasoning) (INT4) | -- | -- | -- | 2.64 | 8.1% | 262k | Open | |||
Gemma 3 4B (FP8) | -- | -- | -- | 2.89 | 13.9% | 131k | Open | |||
Qwen3 Next 80B A3B | -- | -- | -- | 1.01 | 1.8% | 262k | Open | |||
Gemma 3 27B | -- | -- | -- | 12.86 | 0.0% | 131k | Open | |||
Qwen3 235B 2507 | -- | -- | -- | 2.58 | 0.0% | 131k | Open | |||
Qwen3.6 35B A3B (non-reasoning) (FP8) | -- | -- | -- | 1.16 | 1.6% | 262k | Open | |||
Gemma 4 26B A4B | -- | -- | -- | 13.40 | 1.5% | 256k | Open | |||
Gemma 4 26B A4B (non-reasoning) | -- | -- | -- | 1.68 | 1.4% | 262k | Open | |||
Qwen3.6 35B A3B | -- | -- | -- | 25.81 | 2.6% | 262k | Open | |||
DeepSeek V4 Flash (high) (FP8) | -- | -- | -- | -- | -- | 1M | Open | |||
Llama 4 Maverick (FP8) | -- | -- | -- | 1.21 | 5.3% | 1M | Open | |||
Llama 3.3 70B (FP8) | -- | -- | -- | 7.40 | 5.2% | 131k | Open | |||
gpt-oss-120b (low) | -- | -- | -- | 3.27 | 4.5% | 131k | Open | |||
Gemma 4 31B (non-reasoning) | -- | -- | -- | 2.51 | 1.3% | 262k | Open | |||
Qwen3 VL 235B A22B (FP8) | -- | -- | -- | 3.18 | 8.3% | 131k | Open | |||
DeepSeek V4 Flash (max) (FP8) | -- | -- | -- | 8.50 | 5.7% | 1M | Open | |||
DeepSeek V4.1 Flash (max) | -- | -- | -- | 28.52 | 0.0% | 1M | Open | |||
GLM-5.2 (max) (NVFP4) | -- | -- | -- | 16.85 | 5.5% | 1M | Open | |||
Kimi K3 (max) | -- | -- | -- | 29.78 | 0.0% | 1M | Open | |||
Qwen3 Coder Next (FP8) | -- | -- | -- | 1.03 | 6.1% | 262k | Open | |||
DeepSeek V4 Flash 0731 (max) | -- | -- | -- | 14.89 | 0.0% | 1M | Open | |||
Key definitions
Frequently Asked Questions
Common questions about Parasail
Parasail offers 31 models that we track: Kimi K3.1 Pro, DeepSeek V4.1 Flash 1017 (max), GLM-5.3 (max), MiniMax-M3.1-Flash, GLM-5.3-Flash, Qwen3.8 27B (xhigh) (FP8), MiniMax-M3 (MXFP8), Gemma 4 31B, gpt-oss-120b (high), Kimi K2.6, Qwen3.5 397B A17B, Kimi K2.6 (non-reasoning) (INT4), Gemma 3 4B (FP8), Qwen3 Next 80B A3B, Gemma 3 27B, Qwen3 235B 2507, Qwen3.6 35B A3B (non-reasoning) (FP8), Gemma 4 26B A4B, Gemma 4 26B A4B (non-reasoning), Qwen3.6 35B A3B, Llama 4 Maverick (FP8), Llama 3.3 70B (FP8), gpt-oss-120b (low), Gemma 4 31B (non-reasoning), Qwen3 VL 235B A22B (FP8), DeepSeek V4 Flash (max) (FP8), DeepSeek V4.1 Flash (max), GLM-5.2 (max) (NVFP4), Kimi K3 (max), Qwen3 Coder Next (FP8), and DeepSeek V4 Flash 0731 (max).
The most intelligent model available on Parasail is Kimi K3.1 Pro with an Intelligence Index score of 46.
The model with the lowest time per task on Parasail is gpt-oss-120b (high) at 6.8 minutes per Intelligence Index task. Time per task is the output tokens per task divided by end-to-end output speed, so lower means a task finishes sooner.
The model with the lowest time to first answer token on Parasail is Qwen3 Next 80B A3B at 0.99s. Lower latency means faster initial response time.
The most affordable model on Parasail by blended price is Gemma 3 4B (FP8) at $0.045 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on Parasail vary up to 51.3x across models, from $0.045 per 1M tokens for Gemma 3 4B (FP8) to $2.31 per 1M tokens for Kimi K3 (max).
Yes, Parasail offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
Yes, all 31 models on Parasail support JSON mode for structured output.
29 of 31 models on Parasail support function calling (tool use).
Yes, Parasail offers 19 reasoning models: Kimi K3.1 Pro, DeepSeek V4.1 Flash 1017 (max), GLM-5.3 (max), MiniMax-M3.1-Flash, GLM-5.3-Flash, Qwen3.8 27B (xhigh) (FP8), MiniMax-M3 (MXFP8), Gemma 4 31B, gpt-oss-120b (high), Kimi K2.6, Qwen3.5 397B A17B, Gemma 4 26B A4B, Qwen3.6 35B A3B, gpt-oss-120b (low), DeepSeek V4 Flash (max) (FP8), DeepSeek V4.1 Flash (max), GLM-5.2 (max) (NVFP4), Kimi K3 (max), and DeepSeek V4 Flash 0731 (max). Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, all 31 models on Parasail are open weight models.
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 Parasail, consider: intelligence (for quality-sensitive tasks), time per task (how quickly the model completes an Intelligence Index task), 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.