Model comparison
Inkling vs Llama-3.3-70B-Instruct
Inkling is the stronger model overall, scoring 44.1 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 17× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
MetaSummary
- They share 26 benchmarks with published results for both. Inkling scores higher in 9 categories and Llama-3.3-70B-Instruct in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Inkling leads 40.4 to 14.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 88.9% for Inkling and 5.1% for Llama-3.3-70B-Instruct.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Llama-3.3-70B-Instruct accepts more context: 128K tokens versus 66K.
Side by side
| Inkling | Llama-3.3-70B-Instruct | |
|---|---|---|
| Provider | Thinking Machines Lab | Meta |
| Noometry Index | 44.1 | 30.6 |
| Released | 2026-07-15 | 2024-12-06 |
| Weights | Open | Open |
| Context window | 66K | 128K |
| Max output | 66K | 4K |
| Input $ / M tokens | $1.87 | $0.10 |
| Output $ / M tokens | $4.68 | $0.32 |
| Results tracked | 41 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling leads
Inkling: 34.5 (#234), Llama-3.3-70B-Instruct: 31.0 (#290)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| SciCode | 47% | 26% |
| WeirdML | 32.3% | 14.4% |
| LMArena Coding | 1464 | 1268 |
| FrontierCode | 14% | — |
| LMArena WebDev | 1413 | — |
| FrontierSWE | 4.1% | — |
| BigCodeBench Instruct | — | 46.9% |
| LiveBench Coding | — | 36.6% |
| BigCodeBench Complete | — | 57.5% |
| ALE-Bench | 946 | — |
Agentic & Tool Use Inkling leads
Inkling: 29.6 (#85), Llama-3.3-70B-Instruct: 25.8 (#105)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| APEX-Agents | 33.8% | — |
| Berkeley Function Calling Leaderboard | — | 31.9% |
| τ²-bench Banking | 25% | — |
| BALROG | — | 23% |
Reasoning Inkling leads
Inkling: 40.4 (#56), Llama-3.3-70B-Instruct: 14.1 (#327)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| SimpleBench | 50% | 19.9% |
| CritPt | 5.4% | 0% |
| LMArena Hard Prompts | 1451 | 1257 |
| DTBench | 87.5% | 59.5% |
| LMCA | 37.6% | 17.5% |
| Epoch Capabilities Index | 148.54 | 127.33 |
| ARC-AGI-2 | 36.5% | — |
| ARC-AGI-1 | 79.5% | — |
| Chess Puzzles | 21% | — |
| LiveBench Reasoning | — | 50.8% |
| LiveBench Data Analysis | — | 49.5% |
| ForecastBench | — | 58.6 |
| LiveBench | — | 50.2% |
Math Inkling leads
Inkling: 31.3 (#225), Llama-3.3-70B-Instruct: 15.3 (#298)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 5.1% |
| LMArena Math | 1479 | 1267 |
| FrontierMath (Tiers 1-3) | 33.3% | — |
| FrontierMath Tier 4 | 4.9% | — |
| ProofBench | 0% | — |
| LiveBench Math | — | 42.2% |
| MATH Level 5 | — | 41.6% |
Knowledge Inkling leads
Inkling: 55.1 (#49), Llama-3.3-70B-Instruct: 30.6 (#226)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| GPQA Diamond | 88.3% | 47.4% |
| LMArena Expert | 1465 | 1225 |
| SimpleQA Verified | 40.3% | — |
| Confabulations | — | 22.8% |
| Vectara Hallucination Rate | — | 4.1% |
| MMLU | — | 86.3% |
Multilingual Inkling leads
Inkling: 54.0 (#52), Llama-3.3-70B-Instruct: 39.9 (#220)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | 1434 | 1236 |
| LMArena Chinese | 1490 | 1217 |
| LMArena French | 1458 | 1281 |
| LMArena German | 1446 | 1251 |
| LMArena Japanese | 1429 | 1150 |
| LMArena Korean | 1404 | 1143 |
| LMArena Russian | 1429 | 1252 |
| LMArena Spanish | 1448 | 1270 |
Instruction Following Inkling leads
Inkling: 75.1 (#71), Llama-3.3-70B-Instruct: 71.1 (#157)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Instruction Following | 1426 | 1242 |
| LiveBench Instruction Following | — | 82.7% |
Long Context Inkling leads
Inkling: 43.8 (#86), Llama-3.3-70B-Instruct: 26.4 (#295)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Longer Query | 1434 | 1256 |
| Fiction.LiveBench | — | 33.3% |
Writing & Preference Inkling leads
Inkling: 65.2 (#51), Llama-3.3-70B-Instruct: 47.6 (#207)
| Benchmark | Inkling | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | 1441 | 1274 |
| LMArena Creative Writing | 1387 | 1250 |
| LMArena Multi-Turn | 1436 | 1280 |
| EQ-Bench Creative Writing | 1611 | — |
| EQ-Bench 4 | 1226 | — |
| LiveBench Language | — | 39.2% |
Frequently asked questions
Is Inkling better than Llama-3.3-70B-Instruct?
Inkling is the stronger model overall, scoring 44.1 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 17× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, Inkling or Llama-3.3-70B-Instruct?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Inkling or Llama-3.3-70B-Instruct better for coding?
Inkling scores higher on coding benchmarks: 34.5 versus 31.0 in the Noometry coding category.
Which has the bigger context window?
Llama-3.3-70B-Instruct does, with 128K tokens against 66K.
How many benchmarks do Inkling and Llama-3.3-70B-Instruct share?
26 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and Llama-3.3-70B-Instruct has 43.