Granite 4.2 8B vs Hermes 3 70B Instruct
Granite 4.2 8B leads Hermes 3 70B Instruct on the Artificial Analysis Intelligence Index, 11.1 to 6.7, and Granite 4.2 8B costs 3.5x less per output token ($0.20 vs $0.70 per million). Updated October 23, 2026; 3 benchmarks have a result for both.
Granite 4.2 8B or Hermes 3 70B Instruct: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Granite 4.2 8B for output tokens 3.5x cheaper ($0.20 vs $0.70 per million).
- Shared results
- 3
- Granite 4.2 8B only
- 24
- Hermes 3 70B Instruct only
- 2
- Like-for-like categories
- 2 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
Granite 4.2 8B
Granite 4.2 8B leads, 11.1 against 6.7.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Granite 4.2 8B has a AA Agentic Index result, 1.3.
Directional only
Long documents
Listed context window
No call
Both take 131K tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Granite 4.2 8B
Granite 4.2 8B $0.0002, Hermes 3 70B Instruct $0.0010 per turn.
List prices
At a glance
- Coding index
- Granite 4.2 8B not benchmarked, Hermes 3 70B Instruct not benchmarked.
- Widest category
- Reasoning, 14.4 points apart on average over 2 shared benchmarks.
- Widest single gap
- GPQA Diamond (Artificial Analysis run): Granite 4.2 8B 63.1% against Hermes 3 70B Instruct 40.1%, 23.0 points apart on a 0 to 100 scale.
- Output price
- Granite 4.2 8B $0.20 against Hermes 3 70B Instruct $0.70 per million output tokens, 3.5x apart.
- Context window
- Both take 131K tokens.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | Granite 4.2 8B | Bars | Hermes 3 70B Instruct | Ties |
|---|---|---|---|---|
| Intelligence | 1 | 0 | ||
| Reasoning | 2 | 0 | ||
| All shared boards | 3 | 0 |
CodingDirectional onlySix of these boards feed the Vibe Coding Index5 boards, none shared
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| AA Coding Index | 22.4 | Not measured | Only Granite 4.2 8B measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 18.8% | Only Hermes 3 70B Instruct measured |
| SciCode | 31.5% | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: coding | 1378 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: software and IT | 1366 | Not measured | Only Granite 4.2 8B measured |
| Neither model has a result on the other 38 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| AA Intelligence Index | 11.1 | 6.7 | Granite 4.2 8B |
| Neither model has a result on the other 1 board here | |||
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| AA Agentic Index | 1.3 | Not measured | Only Granite 4.2 8B measured |
| Neither model has a result on the other 41 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 63.1% | 40.1% | Granite 4.2 8B |
| Humanity's Last Exam (Artificial Analysis run) | 9.7% | 4.0% | Granite 4.2 8B |
| LMArena Text: hard prompts | 1327 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: expert | 1386 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: hard prompts in English | 1344 | Not measured | Only Granite 4.2 8B measured |
| Neither model has a result on the other 18 boards here | |||
MathNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| Neither model has a result on 14 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | Not measured | 57.1% | Only Hermes 3 70B Instruct measured |
| LMArena Text: science | 1335 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: medicine and healthcare | 1332 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: legal and government | 1327 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: business and finance | 1322 | Not measured | Only Granite 4.2 8B measured |
| Neither model has a result on the other 15 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index12 boards, none shared
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| Long context reasoning (AA-LCR) | 45.0% | Not measured | Only Granite 4.2 8B measured |
| LMArena Text | 1319 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: instruction following | 1301 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: creative writing | 1235 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: English | 1337 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: non-English | 1300 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: Chinese | 1365 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: Russian | 1280 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: multi-turn | 1300 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: longer queries | 1324 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: writing and language | 1269 | Not measured | Only Granite 4.2 8B measured |
| LMArena Text: entertainment and media | 1271 | Not measured | Only Granite 4.2 8B measured |
| Neither model has a result on the other 36 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Granite 4.2 8B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Granite 4.2 8B costs less
Repository review
50K input + 3K output tokens
Granite 4.2 8B costs less
Agent loop
200K input + 10K output tokens
Granite 4.2 8B costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | IBM Granite 4.2 8B | Nous Hermes 3 70B Instruct |
|---|---|---|
| Maker | IBM Granite | Nous Research |
| Weights | Open weights | Open weights |
| Released | August 31, 2026 | August 18, 2024 |
| Context window | 131K | 131K |
| Input $/Mtok | $0.08 | $0.70 |
| Output $/Mtok | $0.20 | $0.70 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Granite 4.2 8B better than Hermes 3 70B Instruct for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Granite 4.2 8B for output tokens 3.5x cheaper ($0.20 vs $0.70 per million).
Which is cheaper, Granite 4.2 8B or Hermes 3 70B Instruct?
Granite 4.2 8B costs $0.08 per million input tokens and $0.20 per million output tokens; Hermes 3 70B Instruct costs $0.70 and $0.70. Cost per solved task is known for neither of them.
Which has the bigger context window, Granite 4.2 8B or Hermes 3 70B Instruct?
Neither: both take 131K tokens.
How many benchmarks is the Granite 4.2 8B vs Hermes 3 70B Instruct comparison based on?
3 benchmarks have a result for both models as of October 23, 2026. Granite 4.2 8B's coding index rests on 0 independent benchmarks and Hermes 3 70B Instruct's on 0. Where only one model has a result, the page says so instead of naming a winner.