Skip to content
OSMEXO
Sign inGet an API key
☰
Sign inGet an API key

Granite 4.2 8B

ibm/granite-4.2-8b
  • Ibm
  • Family · Granite 4
  • Released 25 Aug 2026
  • Open weights
  • Text → Text
  • Reasoning · adjustable
  • Tools · JSON
Not on OSMEXO yet
Request this model
Spec & list price
Context128K
Max output118K
Input / 1M$0.06
Output / 1M$0.152.5× input
Scores · standing
Intelligence11.1#232/490
Coding22.4#135/196
Usage
Tokens / day1.2B▼ 1.9% 7d
Usage rank#188
Requests / day118.7K
Tool calls / req0.05
Per request · 30 days
Avg input4.5K0.3× median
Avg output6200.9× median
Cache hit58%0.8× median
Reasoning share50%0.9× median
Tool-call failures8.9%4.7× median

Consumption

1.7B3.3B5.0B3 Oct17 Oct1 Nov
Peak 4.7B 7 OctLow 325.3M 10 OctRank today #188

Benchmarks

bar = share of the top score · Benchmark data: Artificial Analysis
GPQA Diamond#257 of 445
63.1%
Long-context reasoning#220 of 393
45.0%
SciCode#137 of 160
31.5%
Coding index#135 of 196
22.4
Terminal-Bench 2.1#126 of 182
18.4%
Intelligence index#232 of 490
11.1
Humanity's Last Exam#227 of 465
9.7%
τ-Bench Banking#111 of 154
7.6%
Speed · median across 2 providers, latest day
Output speed#123/33478 tok/s1.2× faster than median
First token#120.28 s3.9× quicker than median
First answer token#7528 sthinking included

#3 cheapest of 33 with a similar score

Value
Intelligence Index
8101214$0.05$0.1$0.2$0.5Granite 4.2 8B
Blended price per 1M tokens (log)
Family progress
Intelligence Index
468101214Nov2026MarMayJulGranite 4.2 8B
  • Family releases11.1

Technical details

Reasoning
on · high
Tokenizer
Other
License
Apache 2.0
Parameters
8.8B
Moderation
none
Downloads · likes
70.4K · 112
Reasoning effort
  • high · default
  • low
  • none
Request parameters 20
  • tools
  • tool_choice
  • response_format
  • structured_outputs
  • reasoning
  • reasoning_effort
  • include_reasoning
  • temperature
  • top_p
  • top_k
+10 more
  • min_p
  • max_tokens
  • stop
  • seed
  • frequency_penalty
  • logit_bias
  • logprobs
  • presence_penalty
  • repetition_penalty
  • top_logprobs

When to use

Best for

  • Private self-hosted assistants
  • Tool calling and structured data extraction
  • Science question answering on a budget
  • Reasoning on or off per request
  • Long document summarization

Think twice for

  • Hardest autonomous coding agents
  • Image or audio input
  • Frontier-level reasoning tasks

Research papers

Same family

Granite 4

Alternatives

other vendors · similar capability

Want Granite 4.2 8B on OSMEXO?

Sign in to request it. Requests are reviewed; a request does not guarantee availability.

Request this model on OSMEXO