Weights & Biases: Models Intelligence, Performance & Price
Analysis of Weights & Biases's models across key metrics including quality, price, output speed, latency, context window & more. This analysis is intended to support you in choosing the best model provided by Weights & Biases for your use-case.
Most Intelligent
GLM-5 (FP8)
Kimi K2.5
MiniMax-M2.5
NVIDIA Nemotron 3 Super
gpt-oss-120B (high)Intelligence index
Total 17 models
Fastest
gpt-oss-20B (low)gpt-oss-20B (high)
Phi-4 MiniNVIDIA Nemotron 3 Super
Llama 3.1 8BOutput speed
Total 17 models
Lowest Price
gpt-oss-20B (high)gpt-oss-20B (low)
Qwen3 235B A22B 2507Qwen3 235B 2507Phi-4 MiniBlended price (per 1M tokens)
Total 17 models
Weights & Biases offers 17 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 Weights & Biases are GLM-5 (FP8) (50), Kimi K2.5 (47), MiniMax-M2.5 (42).
- For output speed, the fastest models are gpt-oss-20B (low) (316 t/s), gpt-oss-20B (high) (296 t/s), Phi-4 Mini (235 t/s). Speed varies significantly across models, with a 121% difference between the fastest and slowest.
- For latency, Phi-4 Mini (0.65s), gpt-oss-20B (low) (0.65s), gpt-oss-20B (high) (0.65s) offer the lowest time to first token.
- For pricing, gpt-oss-20B (high) ($0.09), gpt-oss-20B (low) ($0.09), Qwen3 235B A22B 2507 ($0.10) offer the lowest blended prices per 1M tokens.
- For context window size, Kimi K2.5 (262k), NVIDIA Nemotron 3 Super (262k), Qwen3 235B A22B 2507 (262k) support the largest context windows on Weights & Biases.
Intelligence Evaluations
Artificial Analysis Intelligence Index
Intelligence Evaluations
Intelligence vs. Price
Context Window
Context Window
JSON Mode & Function Calling
Function (Tool) Calling & JSON Mode
| Models | Function calling | JSON Mode |
|---|---|---|
| GLM-5 (FP8), Weights & Biases | ||
| Kimi K2.5, Weights & Biases | ||
| MiniMax-M2.5, Weights & Biases | ||
| NVIDIA Nemotron 3 Super, Weights & Biases | ||
| gpt-oss-120B (high), Weights & Biases | ||
| Qwen3 235B A22B 2507, Weights & Biases | ||
| DeepSeek V3.1, Weights & Biases | ||
| Qwen3 235B 2507, Weights & Biases | ||
| Qwen3 Coder 480B, Weights & Biases | ||
| gpt-oss-120B (low), Weights & Biases | ||
| gpt-oss-20B (high), Weights & Biases | ||
| gpt-oss-20B (low), Weights & Biases |
Pricing
Intelligence 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
Time to First Answer Token
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
Key definitions
Frequently Asked Questions
Common questions about Weights & Biases
Weights & Biases offers 17 models that we track: GLM-5, Kimi K2.5, MiniMax-M2.5, NVIDIA Nemotron 3 Super, gpt-oss-120B (high), Qwen3 235B A22B 2507, DeepSeek V3.1, Qwen3 235B 2507, Qwen3 Coder 480B, gpt-oss-120B (low), gpt-oss-20B (high), gpt-oss-20B (low), Qwen3 30B A3B 2507, Llama 3.3 70B, Llama 4 Scout, Llama 3.1 8B, and Phi-4 Mini.
The most intelligent model available on Weights & Biases is GLM-5 with an Intelligence Index score of 50.
The fastest model on Weights & Biases by output speed is gpt-oss-20B (low) at 316.2 tokens per second.
The model with the lowest time to first token on Weights & Biases is Phi-4 Mini at 0.65s. Lower latency means faster initial response time.
The most affordable model on Weights & Biases by blended price is gpt-oss-20B (high) at $0.09 per 1M tokens (3:1 input to output ratio).
Prices on Weights & Biases vary up to 18x across models, from $0.09 per 1M tokens for gpt-oss-20B (high) to $1.55 per 1M tokens for GLM-5.
Yes, Weights & Biases offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
Yes, all 17 models on Weights & Biases support JSON mode for structured output.
13 of 17 models on Weights & Biases support function calling (tool use).
Yes, Weights & Biases offers 9 reasoning models: GLM-5, Kimi K2.5, MiniMax-M2.5, NVIDIA Nemotron 3 Super, gpt-oss-120B (high), Qwen3 235B A22B 2507, gpt-oss-120B (low), gpt-oss-20B (high), and gpt-oss-20B (low). Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, all 17 models on Weights & Biases 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 Weights & Biases, 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.