IBM: Granite 4.2 8B
by IBM GraniteIBM: Granite 4.2 8B is a large language model from IBM Granite. It costs $0.060 per million input tokens and $0.250 per million output tokens. Its context window is 131K tokens.
- Input / 1M tokens
- $0.060
- Output / 1M tokens
- $0.250
- Cached input / 1M
- $0.015
- Context window
- 131K
On repeated prefixes
tokens
Who serves it cheapest
2 hosts serve Granite 4.2 8B. Same weights, same API - the price difference is pure margin and routing.
| Provider | Input / 1M | Output / 1M | Context | Throughput | Uptime 24h |
|---|---|---|---|---|---|
| DeepInfraCheapestbf16 | $0.060 | $0.250 | 131K | - | 100.0% |
| CoreWeavebf16 | $0.100 | $0.150 | 131K | - | 100.0% |
The spread between DeepInfra and CoreWeave is about the same for identical weights. Quantization and context limits differ, so check both columns before switching.
Benchmarks
Independent scores published alongside the catalogue.
- Intelligence index
- 11.1
- Coding index
- 22.4
- Agentic index
- 1.3
About Granite 4.2 8B
Granite 4.2 8B is a dense reasoning model from IBM. It is suited for mathematics, code generation, multilingual dialogue, and agentic workflows that need multi-step reasoning. It supports full, low-effort,...
Specifications
| Model ID | ibm-granite/granite-4.2-8b |
|---|---|
| Provider | IBM Granite |
| Context window | 131K tokens |
| Max output | 118K tokens |
| Input modalities | text |
| Output modalities | text |
| Knowledge cutoff | - |
| Open weights | Yes - ibm-granite/granite-4.2-8b |
| Released | August 31, 2026 |
Cheaper alternatives
Models that cost less than Granite 4.2 8B while keeping at least half its context window and every input modality it supports.
Mistral Small 3.2 24B
$0.075 in · $0.200 out
about the same cheaperNemotron 3 Nano 30B A3B
$0.060 in · $0.240 out
about the same cheaperNova Lite 1.0
$0.060 in · $0.240 out
about the same cheaperNemotron 3.5 Lightning
$0.070 in · $0.200 out
about the same cheaperMinistral 3 3B 2512
$0.100 in · $0.100 out
1.1× cheaperFrequently asked
How much does IBM: Granite 4.2 8B cost?
$0.060 per million input tokens and $0.250 per million output tokens. Cached input reads cost $0.015 per million tokens.
What is the context window of IBM: Granite 4.2 8B?
131K tokens, with up to 118K tokens of output per request.
Which provider serves IBM: Granite 4.2 8B cheapest?
DeepInfra at $0.060 per million input tokens - about the same cheaper than CoreWeave, the most expensive of the 2 hosts serving it.
Confirm against the source: IBM Granite official pricing.