Model comparison
Grok 4.7 vs Llama 3.1-70B
Grok 4.7 is the stronger model overall, scoring 53.1 to 29.6 on the Noometry Index. Llama 3.1-70B costs 7.5× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.
Last verified . 20 shared benchmarks.
Summary
- They share 20 benchmarks with published results for both. Grok 4.7 scores higher in 9 categories and Llama 3.1-70B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Grok 4.7 leads 57.8 to 13.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.1% for Grok 4.7 and 3.6% for Llama 3.1-70B.
- Llama 3.1-70B is cheaper at $0.40 / $0.40 per million input/output tokens, against $2 / $6 for Grok 4.7.
- Grok 4.7 accepts more context: 500K tokens versus 128K.
- Llama 3.1-70B has downloadable open weights; the other is API-only.
Side by side
| Grok 4.7 | Llama 3.1-70B | |
|---|---|---|
| Provider | xAI | Meta |
| Noometry Index | 53.1 | 29.6 |
| Released | 2026-09-21 | 2024-07-23 |
| Weights | Proprietary | Open |
| Context window | 500K | 128K |
| Max output | 500K | 4K |
| Input $ / M tokens | $2 | $0.40 |
| Output $ / M tokens | $6 | $0.40 |
| Results tracked | 39 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.7 leads
Grok 4.7: 58.0 (#18), Llama 3.1-70B: 30.3 (#296)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| LMArena Coding | 1427 | 1260 |
| FrontierCode | 47.6% | — |
| CursorBench | 46.3% | — |
| LMArena WebDev | 1639 | — |
| FrontierSWE | 29.5% | — |
| SciCode | 57.8% | — |
| WeirdML | — | 9% |
| BigCodeBench Instruct | — | 46.1% |
| BigCodeBench Complete | — | 54.8% |
Agentic & Tool Use Grok 4.7 leads
Grok 4.7: 36.7 (#37), Llama 3.1-70B: 25.1 (#112)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| APEX-Agents | 54.6% | — |
| TheAgentCompany | — | 6.9% |
| BALROG | — | 27.9% |
| GDP.pdf | 22.8% | — |
| Vending-Bench 2 | 10,537 | — |
Reasoning Grok 4.7 leads
Grok 4.7: 49.1 (#40), Llama 3.1-70B: 21.6 (#220)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| LMArena Hard Prompts | 1413 | 1241 |
| DTBench | 96% | 60% |
| LMCA | 49.4% | 14.8% |
| Epoch Capabilities Index | 153.53 | 125.92 |
| NYT Connections (extended) | 76.8% | — |
| CritPt | 18% | — |
| Chess Puzzles | 38% | — |
| Mystery Game Puzzles | 29% | — |
Math Grok 4.7 leads
Grok 4.7: 57.8 (#39), Llama 3.1-70B: 13.5 (#304)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.1% | 3.6% |
| LMArena Math | 1407 | 1252 |
| FrontierMath (Tiers 1-3) | 53% | — |
| FrontierMath Tier 4 | 17.1% | — |
| ProofBench | 34% | — |
| Omni-MATH | — | 21% |
| MATH Level 5 | — | 36.7% |
Knowledge Grok 4.7 leads
Grok 4.7: 62.8 (#22), Llama 3.1-70B: 24.2 (#269)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| GPQA Diamond | 92.7% | 44.2% |
| LMArena Expert | 1422 | 1209 |
| SimpleQA Verified | 56% | — |
| MMLU-Pro | — | 65.3% |
| GPQA (HELM) | — | 42.6% |
| MMLU | — | 80.1% |
Multimodal Not comparable
Grok 4.7: 35.5 (#87), Llama 3.1-70B: —
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| LMArena Vision | 1228 | — |
| Blueprint-Bench 2 | 32.5% | — |
| Furniture Assembly | 20.8% | — |
Multilingual Grok 4.7 leads
Grok 4.7: 50.8 (#116), Llama 3.1-70B: 38.8 (#225)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| LMArena Non-English | 1389 | 1219 |
| LMArena Chinese | 1455 | 1215 |
| LMArena French | 1455 | 1261 |
| LMArena Russian | 1397 | 1234 |
| LMArena Spanish | 1400 | 1253 |
| LMArena German | — | 1222 |
| LMArena Japanese | — | 1132 |
| LMArena Korean | — | 1140 |
Instruction Following Grok 4.7 leads
Grok 4.7: 74.1 (#105), Llama 3.1-70B: 65.3 (#223)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| LMArena Instruction Following | 1404 | 1231 |
| IFEval | — | 82.1% |
Long Context Grok 4.7 leads
Grok 4.7: 43.1 (#104), Llama 3.1-70B: 37.6 (#214)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| LMArena Longer Query | 1413 | 1241 |
Writing & Preference Grok 4.7 leads
Grok 4.7: 70.0 (#24), Llama 3.1-70B: 35.4 (#267)
| Benchmark | Grok 4.7 | Llama 3.1-70B |
|---|---|---|
| LMArena Text | 1399 | 1261 |
| LMArena Creative Writing | 1391 | 1232 |
| EQ-Bench Creative Writing | 2007 | 784 |
| LMArena Multi-Turn | 1393 | 1256 |
| WildBench | — | 75.8% |
Frequently asked questions
Is Grok 4.7 better than Llama 3.1-70B?
Grok 4.7 is the stronger model overall, scoring 53.1 to 29.6 on the Noometry Index. Llama 3.1-70B costs 7.5× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.
Which is cheaper, Grok 4.7 or Llama 3.1-70B?
Llama 3.1-70B is cheaper. It lists at $0.40 per million input tokens and $0.40 per million output tokens; Grok 4.7 lists at $2 and $6.
Is Grok 4.7 or Llama 3.1-70B better for coding?
Grok 4.7 scores higher on coding benchmarks: 58.0 versus 30.3 in the Noometry coding category.
Which has the bigger context window?
Grok 4.7 does, with 500K tokens against 128K.
How many benchmarks do Grok 4.7 and Llama 3.1-70B share?
20 benchmarks have published results for both models. Grok 4.7 has 39 scored results on Noometry and Llama 3.1-70B has 35.