Model comparison
GPT-3.5-turbo vs Inkling
Inkling is the stronger model overall, scoring 44.1 to 23.2 on the Noometry Index. GPT-3.5-turbo costs 3.4× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
Summary
- They share 26 benchmarks with published results for both. GPT-3.5-turbo scores higher in 0 categories and Inkling in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Inkling leads 55.1 to 10.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 2.2% for GPT-3.5-turbo and 88.9% for Inkling.
- GPT-3.5-turbo is cheaper at $0.50 / $1.50 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Inkling accepts more context: 66K tokens versus 16K.
- Inkling has downloadable open weights; the other is API-only.
Side by side
| GPT-3.5-turbo | Inkling | |
|---|---|---|
| Provider | OpenAI | Thinking Machines Lab |
| Noometry Index | 23.2 | 44.1 |
| Released | 2023-03-01 | 2026-07-15 |
| Weights | Proprietary | Open |
| Context window | 16K | 66K |
| Max output | 4K | 66K |
| Input $ / M tokens | $0.50 | $1.87 |
| Output $ / M tokens | $1.50 | $4.68 |
| Results tracked | 44 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Inkling leads
GPT-3.5-turbo: 23.9 (#331), Inkling: 34.5 (#234)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| WeirdML | 3.5% | 32.3% |
| LMArena Coding | 1136 | 1464 |
| FrontierCode | — | 14% |
| LMArena WebDev | — | 1413 |
| FrontierSWE | — | 4.1% |
| SciCode | — | 47% |
| BigCodeBench Instruct | 39.1% | — |
| BigCodeBench Complete | 50.6% | — |
| ALE-Bench | — | 946 |
| HumanEval+ | 70.7% | — |
| MBPP+ | 69.7% | — |
Agentic & Tool Use Not comparable
GPT-3.5-turbo: —, Inkling: 29.6 (#85)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| APEX-Agents | — | 33.8% |
| τ²-bench Banking | — | 25% |
| METR Time Horizons | 21.5% | — |
Reasoning Inkling leads
GPT-3.5-turbo: 13.8 (#332), Inkling: 40.4 (#56)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| Chess Puzzles | 0% | 21% |
| LMArena Hard Prompts | 1108 | 1451 |
| DTBench | 48.5% | 87.5% |
| LMCA | 9.7% | 37.6% |
| Epoch Capabilities Index | 118.55 | 148.54 |
| ARC-AGI-2 | — | 36.5% |
| SimpleBench | — | 50% |
| ARC-AGI-1 | — | 79.5% |
| CritPt | — | 5.4% |
| Mystery Game Puzzles | 3% | — |
| Adversarial NLI | 58.1% | — |
| BIG-Bench Hard | 61.6% | — |
| CommonsenseQA 2.0 | 57% | — |
| ForecastBench | 50.4 | — |
| WinoGrande | 81.6% | — |
Math Inkling leads
GPT-3.5-turbo: 6.3 (#327), Inkling: 31.3 (#225)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0% | 33.3% |
| OTIS Mock AIME 2024-2025 | 2.2% | 88.9% |
| LMArena Math | 1142 | 1479 |
| FrontierMath Tier 4 | — | 4.9% |
| ProofBench | — | 0% |
| MATH Level 5 | 15.9% | — |
| GSM8K | 57.8% | — |
Knowledge Inkling leads
GPT-3.5-turbo: 10.0 (#303), Inkling: 55.1 (#49)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| GPQA Diamond | 28% | 88.3% |
| LMArena Expert | 1070 | 1465 |
| SimpleQA Verified | — | 40.3% |
| ARC (AI2) Challenge | 87.4% | — |
| BoolQ | 87% | — |
| MMLU | 71.4% | — |
| OpenBookQA | 86% | — |
| TriviaQA | 85.8% | — |
Multilingual Inkling leads
GPT-3.5-turbo: 31.5 (#258), Inkling: 54.0 (#52)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| LMArena Non-English | 1108 | 1434 |
| LMArena Chinese | 1075 | 1490 |
| LMArena French | 1118 | 1458 |
| LMArena German | 1090 | 1446 |
| LMArena Japanese | 1043 | 1429 |
| LMArena Korean | 1019 | 1404 |
| LMArena Russian | 1123 | 1429 |
| LMArena Spanish | 1121 | 1448 |
Instruction Following Inkling leads
GPT-3.5-turbo: 57.9 (#262), Inkling: 75.1 (#71)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| LMArena Instruction Following | 1119 | 1426 |
Long Context Inkling leads
GPT-3.5-turbo: 34.0 (#254), Inkling: 43.8 (#86)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| LMArena Longer Query | 1121 | 1434 |
Writing & Preference Inkling leads
GPT-3.5-turbo: 25.3 (#305), Inkling: 65.2 (#51)
| Benchmark | GPT-3.5-turbo | Inkling |
|---|---|---|
| LMArena Text | 1125 | 1441 |
| LMArena Creative Writing | 1092 | 1387 |
| EQ-Bench Creative Writing | 451 | 1611 |
| LMArena Multi-Turn | 1117 | 1436 |
| EQ-Bench 4 | — | 1226 |
Frequently asked questions
Is GPT-3.5-turbo better than Inkling?
Inkling is the stronger model overall, scoring 44.1 to 23.2 on the Noometry Index. GPT-3.5-turbo costs 3.4× less per token, which makes it the better buy when Inkling's lead doesn't matter for your workload.
Which is cheaper, GPT-3.5-turbo or Inkling?
GPT-3.5-turbo is cheaper. It lists at $0.50 per million input tokens and $1.50 per million output tokens; Inkling lists at $1.87 and $4.68.
Is GPT-3.5-turbo or Inkling better for coding?
Inkling scores higher on coding benchmarks: 34.5 versus 23.9 in the Noometry coding category.
Which has the bigger context window?
Inkling does, with 66K tokens against 16K.
How many benchmarks do GPT-3.5-turbo and Inkling share?
26 benchmarks have published results for both models. GPT-3.5-turbo has 44 scored results on Noometry and Inkling has 41.