llama-3-2-90b-vision-instruct
watsonx · chat model
Input
$2.0000 / 1M tokens
Output
$2.0000 / 1M tokens
Cached input
N/A
Context
128.0K
Pricing
| Item | Raw value (per token) | Normalized |
|---|---|---|
| Input | 0.000002 | $2.0000 / 1M tokens |
| Output | 0.000002 | $2.0000 / 1M tokens |
| Embedding | 0.000002 | $2.0000 / 1M tokens |
Token limits
Context window
128.0K
Max input tokens
128.0K
Max output tokens
128.0K
Max tokens
128.0K
Capabilities
| Capability | Supported |
|---|---|
| Vision | Yes |
| Function calling | Yes |
| Parallel function calling | Yes |
| Tool choice | No |
| Prompt caching | No |
| Reasoning | No |
| Response schema | No |
| System messages | No |
| Audio input | No |
| Audio output | No |
| Web search | No |
| PDF input | No |
| Video input | No |
Similar models
Models with comparable pricing in the same mode (chat).
| Model | Provider | Input | Output | Context | Coding | Features |
|---|---|---|---|---|---|---|
| Qwen3-235B-A22B | hyperbolic | $2.0000 / 1M tokens | $2.0000 / 1M tokens | 131.1K | N/A | Tools |
| Kimi-K2-Instruct | hyperbolic | $2.0000 / 1M tokens | $2.0000 / 1M tokens | 131.1K | N/A | Tools |
| meta.llama3-2-90b-instruct-v1:0 | Bedrock | $2.0000 / 1M tokens | $2.0000 / 1M tokens | 4.1K | N/A | Vision Tools |
| meta.llama-3.2-90b-vision-instruct | oci | $2.0000 / 1M tokens | $2.0000 / 1M tokens | 4.0K | N/A | Vision Tools |
| meta.llama-3.2-11b-vision-instruct | oci | $2.0000 / 1M tokens | $2.0000 / 1M tokens | 4.0K | N/A | Vision Tools |
| Qwen3-Coder-480B-A35B-Instruct-FP8 | Together AI | $2.0000 / 1M tokens | $2.0000 / 1M tokens | N/A | N/A | Tools |
Sources
| Pricing source | LiteLLM model cost map |
| Synced at | 2026-05-27 |
| Manual review | Not reviewed |
Raw LiteLLM fields
{
"max_tokens": 128000,
"max_input_tokens": 128000,
"max_output_tokens": 128000,
"input_cost_per_token": 0.000002,
"output_cost_per_token": 0.000002,
"litellm_provider": "watsonx",
"mode": "chat",
"supports_function_calling": true,
"supports_parallel_function_calling": true,
"supports_vision": true
}