Gemma 3 27B vs Hermes 3 70B Instruct
Hermes 3 70B Instruct leads Gemma 3 27B on the Artificial Analysis Intelligence Index, 6.7 to 4.9, and Gemma 3 27B costs 2.3x less per output token ($0.30 vs $0.70 per million). Updated October 22, 2026; 5 benchmarks have a result for both.
Gemma 3 27B or Hermes 3 70B Instruct: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Gemma 3 27B for output tokens 2.3x cheaper ($0.30 vs $0.70 per million) and work measured by LiveCodeBench (Artificial Analysis run) or AA Intelligence Index, where it leads.
- Shared results
- 5
- Gemma 3 27B only
- 61
- Hermes 3 70B Instruct only
- 0
- Like-for-like categories
- 4 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
Hermes 3 70B Instruct
Hermes 3 70B Instruct leads, 6.7 against 4.9.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Gemma 3 27B has a AA Agentic Index result, 0.1.
Directional only
Long documents
Listed context window
No call
Both take 131K tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Gemma 3 27B
Gemma 3 27B $0.0002, Hermes 3 70B Instruct $0.0010 per turn.
List prices
At a glance
- Coding index
- Gemma 3 27B not benchmarked, Hermes 3 70B Instruct not benchmarked.
- Widest category
- Knowledge, 9.8 points apart on average over 1 shared benchmark.
- Widest single gap
- MMLU Pro (Artificial Analysis run): Gemma 3 27B 66.9% against Hermes 3 70B Instruct 57.1%, 9.8 points apart on a 0 to 100 scale.
- Output price
- Gemma 3 27B $0.30 against Hermes 3 70B Instruct $0.70 per million output tokens, 2.3x 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 | Gemma 3 27B | Bars | Hermes 3 70B Instruct | Ties |
|---|---|---|---|---|
| Coding | 0 | 1 | ||
| Intelligence | 0 | 1 | ||
| Reasoning | 2 | 0 | ||
| Knowledge | 1 | 0 | ||
| All shared boards | 3 | 2 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index01
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| Aider polyglot | 4.9% | Not measured | Only Gemma 3 27B measured |
| AA Coding Index | 10.1 | Not measured | Only Gemma 3 27B measured |
| LiveCodeBench (Artificial Analysis run) | 13.7% | 18.8% | Hermes 3 70B Instruct |
| SciCode | 23.3% | Not measured | Only Gemma 3 27B measured |
| Terminal-Bench Hard (Artificial Analysis run) | 3.8% | Not measured | Only Gemma 3 27B measured |
| LMArena Text: coding | 1322 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: software and IT | 1354 | Not measured | Only Gemma 3 27B measured |
| SciCode | 0.0% | Not measured | Only Gemma 3 27B measured |
| Neither model has a result on the other 35 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| AA Intelligence Index | 4.9 | 6.7 | Hermes 3 70B Instruct |
| Epoch Capabilities Index | 130.0 | Not measured | Only Gemma 3 27B measured |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index3 boards, none shared
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| AA Agentic Index | 0.1 | Not measured | Only Gemma 3 27B measured |
| τ²-Bench Telecom | 10.5% | Not measured | Only Gemma 3 27B measured |
| Berkeley Function Calling Leaderboard (BFCL) | 29.5% | Not measured | Only Gemma 3 27B measured |
| Neither model has a result on the other 39 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 42.8% | 40.1% | Gemma 3 27B |
| Humanity's Last Exam (Artificial Analysis run) | 4.4% | 4.0% | Gemma 3 27B |
| GPQA Diamond (Epoch AI run) | 47.7% | Not measured | Only Gemma 3 27B measured |
| LMArena Text: hard prompts | 1340 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: expert | 1304 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: hard prompts in English | 1353 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: diagrams | 1156 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: homework | 1160 | Not measured | Only Gemma 3 27B measured |
| Kagi LLM Benchmark | 35.1% | Not measured | Only Gemma 3 27B measured |
| Neither model has a result on the other 14 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| AA Math Index | 20.7 | Not measured | Only Gemma 3 27B measured |
| AIME 2025 (Artificial Analysis run) | 20.7% | Not measured | Only Gemma 3 27B measured |
| OTIS Mock AIME | 22.5% | Not measured | Only Gemma 3 27B measured |
| MATH level 5 | 74.0% | Not measured | Only Gemma 3 27B measured |
| LMArena Text: math | 1312 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: mathematical work | 1327 | Not measured | Only Gemma 3 27B measured |
| Neither model has a result on the other 8 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | 66.9% | 57.1% | Gemma 3 27B |
| GeoBench | 52.0% | Not measured | Only Gemma 3 27B measured |
| LMArena Text: science | 1383 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: medicine and healthcare | 1391 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: legal and government | 1375 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: business and finance | 1360 | Not measured | Only Gemma 3 27B measured |
| FACTS Benchmark Suite | 29.1% | Not measured | Only Gemma 3 27B measured |
| Neither model has a result on the other 13 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index31 boards, none shared
| Benchmark | Gemma 3 27B | Hermes 3 70B Instruct | Leader |
|---|---|---|---|
| IFBench | 31.8% | Not measured | Only Gemma 3 27B measured |
| Long context reasoning (AA-LCR) | 7.3% | Not measured | Only Gemma 3 27B measured |
| DTBench | 52.5% | Not measured | Only Gemma 3 27B measured |
| LMCA | 12.3% | Not measured | Only Gemma 3 27B measured |
| Fiction.LiveBench | 33.3% | Not measured | Only Gemma 3 27B measured |
| Creative writing (Lech Mazur) | 79.9% | Not measured | Only Gemma 3 27B measured |
| LMArena Text | 1358 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: instruction following | 1321 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: creative writing | 1345 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision | 1165 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: English | 1377 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: non-English | 1334 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: Chinese | 1346 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: German | 1362 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: Japanese | 1287 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: Korean | 1308 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: French | 1368 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: Spanish | 1349 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: Russian | 1349 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: Polish | 1345 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: multi-turn | 1345 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: longer queries | 1333 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: writing and language | 1347 | Not measured | Only Gemma 3 27B measured |
| LMArena Text: entertainment and media | 1320 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: English | 1188 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: Chinese | 1164 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: reading text | 1156 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: humor | 1166 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: captioning | 1151 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: recognizing things | 1196 | Not measured | Only Gemma 3 27B measured |
| LMArena Vision: creative writing | 1177 | Not measured | Only Gemma 3 27B measured |
| Neither model has a result on the other 17 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Gemma 3 27B | 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
Gemma 3 27B costs less
Repository review
50K input + 3K output tokens
Gemma 3 27B costs less
Agent loop
200K input + 10K output tokens
Gemma 3 27B 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 | Google Gemma 3 27B | Nous Hermes 3 70B Instruct |
|---|---|---|
| Maker | Nous Research | |
| Weights | Open weights | Open weights |
| Released | March 12, 2025 | August 18, 2024 |
| Context window | 131K | 131K |
| Input $/Mtok | $0.08 | $0.70 |
| Output $/Mtok | $0.30 | $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 Gemma 3 27B better than Hermes 3 70B Instruct for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Gemma 3 27B for output tokens 2.3x cheaper ($0.30 vs $0.70 per million) and work measured by LiveCodeBench (Artificial Analysis run) or AA Intelligence Index, where it leads.
Which is cheaper, Gemma 3 27B or Hermes 3 70B Instruct?
Gemma 3 27B costs $0.08 per million input tokens and $0.30 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, Gemma 3 27B or Hermes 3 70B Instruct?
Neither: both take 131K tokens.
How many benchmarks is the Gemma 3 27B vs Hermes 3 70B Instruct comparison based on?
5 benchmarks have a result for both models as of October 22, 2026. Gemma 3 27B'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.