IBMGranite 4.2 8B
Granite 4.2 8B is an IBM open-weights reasoning model, released 25 Aug 2026. Its best configuration ranks #496 of 785 on the BenchLeader Index at 46.1 ±3.6, in the lower half. The ± is the point: 284 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in knowledge (54) and lowest in agents & tools (35). At $0.108 per million tokens blended it is among the cheapest fifth of ranked models. Output speed of 97 tokens per second puts it faster than most, with a first answer in 21.2 s. Last measured 22 Oct 2026.
- Blended price
- $0.108/M
- $0.060 in · $0.250 out
- Output speed
- 97 tok/s
- measured by Artificial Analysis
- First answer
- 21 s
- first token 0.88 s
- Context
- 131k
- Full answer
- 26 s
- median, reasoning included
- Cost per run
- $0.024
- one full Intelligence Index run
- Released
- 25 Aug 2026
- Overall index46
- Reasoning43
- Coding49
- Agents & tools35
- Knowledge54
- Human preference44
- Long context47
- Composite42
Versions
IBM has shipped 6 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Granite 4.2 8Bthis page | 25 Aug 2026 | 46.1 | #496 |
| Granite 4.1 8B | 29 Apr 2026 | 40.6 | #653 |
| Granite 3.3 8Bno reasoning | 16 Apr 2025 | 39.2 | #695 |
| Granite 3.1 8B | – | 41.8 | #618 |
| Granite 3.0 8B | – | – | – |
| Granite 3.2 8B | – | – | – |
How its index has moved
9 Sept 2026 to 23 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Metadata
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA Openness Indexnot in index | 72.2#21 | Artificial Analysis |
Reasoning
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Hard Prompts | 1306#256 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 63.1%#357 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 9.7%#330 | Artificial Analysis | |
| CritPt | 0.3%#271 | Artificial Analysis |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Coding | 1370#223 | Arena (LMArena) | |
| SciCode (AA)not in index | 31.5%#209 | Artificial Analysis | |
| Terminal-Bench 4.0 (AA)not in index | 0.0%#152 | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 18.4%#178 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| GDPval-AA v2.1not in index | 0.0%#124 | Artificial Analysis | |
| τ³-Banking (AA) | 7.6%#160 | Artificial Analysis | |
| AutomationBenchnot in index | 2.1%#182 | Artificial Analysis | |
| GDP.pdf (AA)not in index | 0.6%#215 | Artificial Analysis | |
| AA-Briefcase v1.1not in index | 310#56 | Artificial Analysis |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-Omniscience | -17.2#231 | Artificial Analysis | |
| AA-Omniscience: accuracynot in index | 11.2%#483 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 68.0%#52 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Instruction Followingnot in index | 1283#254 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1290#261 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1199#292 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1272#267 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1307#251 | Arena (LMArena) |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-LCR | 45.0%#338 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA Intelligence Index v4.3.2 | 11.1#364 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| CoreWeave | 62 tok/s | 0.40 s | $0.100 | $0.150 | 131k | bf16 |
| DeepInfra | 40 tok/s | 0.88 s | $0.060 | $0.250 | 131k | bf16 |
Price history
The price above, per 1M tokens, every day it changed.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0001 | – | 24.3 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0009 | – | 27.4 s |
| Code edit | 6,000 / 1,500 | $0.0007 | – | 36.6 s |
| Agentic coding session | 60,000 / 4,000 | $0.0046 | – | 1.0 min |
| Structured extraction | 2,000 / 200 | $0.0002 | – | 23.3 s |
More from IBM
See also
Questions people ask
- What are Granite 4.2 8B's benchmark scores?
- As of 24 Oct 2026, Granite 4.2 8B scores 46.1 ±3.6 on the BenchLeader Index, strongest in knowledge (54) and weakest in agents & tools (35). That rests on 8 independent benchmark results, and every score on this page links to its publisher.
- How does Granite 4.2 8B rank?
- Granite 4.2 8B is #496 of 785 ranked configurations on 24 Oct 2026. 284 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Granite 4.2 8B cost per 1M tokens?
- As of 24 Oct 2026, Granite 4.2 8B costs $0.060 per 1M input tokens and $0.250 per 1M output tokens, $0.108 blended (three input tokens per output token). The list price last changed on 9 Sept 2026.
- When was Granite 4.2 8B released?
- IBM released Granite 4.2 8B on 25 Aug 2026.
Data as of 24 Oct 2026. Compare with another model.
Cite as: BenchLeader, “Granite 4.2 8B: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/granite-4-2-8b.