Grok 4.20 vs Inkling 2 Small
Full profiles: Grok 4.20 (SpaceXAI) and Inkling 2 Small (Thinking Machines): every benchmark, price and rank.
Updated · Data from 11 sources
As of 16 Nov 2026, Grok 4.20 and Inkling 2 Small are neck and neck on quality (58.3 vs 57.9 on the BenchLeader Index, inside the ±5.4 margin), and Inkling 2 Small is 3.3× cheaper at $0.475 vs $1.56 per 1M tokens blended. Choose Grok 4.20 for reasoning (61 vs 53) and human preference (66 vs 59) and fast first answers (12.1 s vs 49.0 s); choose Inkling 2 Small for agents & tools (61 vs 51), long context (66 vs 59), and coding (60 vs 53).
Key facts (as of 16 Nov 2026)
- Grok 4.20 scores 58.3 on the BenchLeader Index and Inkling 2 Small scores 57.9 (BenchLeader, 16 Nov 2026).
- Grok 4.20 costs $1.56 per 1M tokens blended (list price, 16 Nov 2026).
- Inkling 2 Small costs $0.475 per 1M tokens blended (list price, 16 Nov 2026).
- Of the 16 benchmarks both models have, Inkling 2 Small leads 10 (BenchLeader, 16 Nov 2026).
- On Arena WebDev, Grok 4.20 is rated 1376 and Inkling 2 Small 1647 (LMArena, as published by 16 Nov 2026).
Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Inkling 2 Small | 60 vs 53 category score; Arena WebDev 1647 vs 1376 | Medium |
| Agents & tool use | Inkling 2 Small | 61 vs 51 category score; Legal Research Bench 46.6% vs 13.9% | High |
| Reasoning | Grok 4.20 | 61 vs 53 category score; Arena Hard Prompts 1489 vs 1438 | High |
| Knowledge | Either | 53 vs 54 category score | Either |
| Long documents | Inkling 2 Small | 66 vs 59 category score; AA-LCR v1.1 83.3% vs 69.0% | High |
| Images and other media | Inkling 2 Small | 64 vs 58 category score; MMMU-Pro 81.6% vs 74.6% | Medium |
| Chat quality (human preference) | Grok 4.20 | 66 vs 59 on Arena votes | Medium |
| Lowest cost | Inkling 2 Small | $0.475 vs $1.56 per 1M blended; agentic coding session $0.022 vs $0.085 | – |
| Fastest first answer | Grok 4.20 | 12.1 s (Grok 4.20 (thinking)) vs 49.0 s (Inkling 2 Small) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Grok 4.20 (thinking) and Inkling 2 Small are neck and neck on quality (58.3 vs 57.9, within the ±5.4 margin).
- Grok 4.20 (thinking) is stronger in human preference, reasoning.
- Inkling 2 Small is stronger in agents & tools, coding, knowledge, long context, multimodal.
- Inkling 2 Small is 3.3× cheaper ($0.475 vs $1.56 per 1M blended).
At a glance
| Spec | Grok 4.20 | Inkling 2 Small |
|---|---|---|
| Developer | SpaceXAI | Thinking Machines |
| Released | 17 Feb 2026 | 5 Nov 2026 |
| Context window (tokens) | 2,000,000 | 1,048,576 |
| Compared at | Grok 4.20 (thinking) | Inkling 2 Small |
| Open weights | No | No |
| Price in / out / cached input per 1M | $1.25 / $2.50 / $0.200 | $0.300 / $1.00 / $0.030 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Grok 4.20 (thinking) | Inkling 2 Small | Leader |
|---|---|---|---|
| Reasoning | 60.5 | 52.8 | Grok 4.20 |
| Coding | 53.3 | 60.3 | Inkling 2 Small |
| Agents & tools | 50.5 | 61.1 | Inkling 2 Small |
| Maths | 53.7 | – | not comparable |
| Knowledge | 53.0 | 53.5 | level |
| Instruction following | 78.7 | – | not comparable |
| Human preference | 66.0 | 58.6 | Grok 4.20 |
| Multimodal | 58.3 | 64.0 | Inkling 2 Small |
| Long context | 58.9 | 66.4 | Inkling 2 Small |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Maths, Instruction following.
Where sources disagree
Across the 16 benchmarks both have, Grok 4.20 leads 6, Inkling 2 Small leads 10. Every overall index that rates both puts Grok 4.20 ahead.
| Overall index | Grok 4.20 (thinking) | Inkling 2 Small | Leader |
|---|---|---|---|
| BenchLeader Index | 58.3 | 57.9 | neck and neck |
| Arena Text | 1472 | 1412 | Grok 4.20 |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | Grok 4.20 (thinking) | Inkling 2 Small |
|---|---|---|
| Blended price per 1M tokens | $1.56 | $0.475 |
| Input / output per 1M | $1.25 / $2.50 | $0.300 / $1.00 |
| Cached input per 1M | $0.200 | $0.030 |
| Price per task: chat reply | $0.0013 | $0.0004 |
| Price per task: summarize a 30-page report | $0.017 | $0.0042 |
| Price per task: code edit | $0.011 | $0.0033 |
| Price per task: agentic coding session | $0.085 | $0.022 |
| Price per task: structured extraction | $0.0030 | $0.0008 |
| Agentic coding session, input mostly cached | $0.038 | $0.0099 |
| Cost per benchmark task (AA Index) | – | $1.18 |
| Output tokens per task (AA Index) | – | 150k |
| Of which reasoning | – | 88k |
| Time per task (AA Index) | – | 25 min |
| Output speed | 111 tok/s | 100 tok/s |
| Time to first answer | 12.1 s | 49.0 s |
| Time to a complete answer, last measured 13 Nov 2026 | 43.8 s | 163.6 s |
Each model is compared at its best-ranked reasoning-effort setting. Grok 4.20 (thinking) scores 58.3 and takes 12 s to its first answer token. Inkling 2 Small scores 57.9 and takes 49 s to its first answer token. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
128 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | Grok 4.20 (thinking) | Inkling 2 Small |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 58.3 | 57.9 |
| Agents & tools score | BenchLeader | 50.5 | 61.1 |
| Coding score | BenchLeader | 53.3 | 60.3 |
| Human preference score | BenchLeader | 66.0 | 58.6 |
| Instruction following score | BenchLeader | 78.7 | – |
| Knowledge score | BenchLeader | 53.0 | 53.5 |
| Long context score | BenchLeader | 58.9 | 66.4 |
| Maths score | BenchLeader | 53.7 | – |
| Multimodal score | BenchLeader | 58.3 | 64.0 |
| Reasoning score | BenchLeader | 60.5 | 52.8 |
| Blended price $/M | BenchLeader | $1.56 | $0.475 |
| Output speed | BenchLeader | 111 tok/s | 100 tok/s |
| Time to first answer | BenchLeader | 12.1 s | 49.0 s |
| Context window | BenchLeader | 2,000,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 89.3% | – |
| Arena Hard Prompts | Arena (LMArena) | 1489 | 1438 |
| GPQA Diamond (AA) | Artificial Analysis | 91.1% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 34.5% | 44.1% |
| HLE-Diamond (AA) | Artificial Analysis | – | 28.7% |
| GPQA Diamond (Vals) | Vals AI | 88.6% | – |
| Kagi LLM Benchmark | Kagi LLM Benchmark | 75.0% | – |
| ARC-AGI-1 | ARC Prize | 89.5% | 96.5%† |
| ARC-AGI-2 | ARC Prize | 65.1% | 81.7%† |
| ARC-AGI-3 | ARC Prize | 0.1% | – |
| CritPt | Artificial Analysis | 6.6% | – |
| CritPt v2 hard subset (AA) | Artificial Analysis | – | 35.0% |
| MysteryMechanism | Vals AI | – | 25.2% |
| Arena Expert | Arena (LMArena) | 1477 | 1488 |
| Chess Puzzles | Epoch AI Benchmarking Hub | 29.0% | – |
| CL-bench | Epoch AI Benchmarking Hub | 22.2%† | – |
| CL-bench Life | Epoch AI Benchmarking Hub | 11.9%† | – |
| LMCA | Epoch AI Benchmarking Hub | 38.7%† | – |
| DTBench | Epoch AI Benchmarking Hub | 90.1%† | – |
| ForecastBench | ForecastBench | 64.1% | – |
| Coding | |||
| SciCode | SciCode | – | 52.3% |
| WeirdML | WeirdML | 52.3%† | – |
| Arena Coding | Arena (LMArena) | 1511 | 1485 |
| Arena WebDev | Arena (LMArena) | 1376 | 1647 |
| LiveCodeBench | Vals AI | 84.3% | – |
Show all 128 benchmarks
| Metric | Publisher | Grok 4.20 (thinking) | Inkling 2 Small |
|---|---|---|---|
| Coding | |||
| SWE-bench (Vals) | Vals AI | 72.2% | – |
| Code Migration | Vals AI | 0.3% | – |
| ProgramBench | Vals AI | – | 0.5% |
| SWE-bench Multimodal | Vals AI | – | 43.1% |
| Vibe Code Bench v1.1 | Vals AI | 4.1% | 81.5% |
| SWE Atlas v1.1 | Scale AI SEAL | – | 61.9% |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | – | 74.2% |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | – | 58.1% |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | – | 53.3% |
| SWE-Bench Pro V2 | Scale AI SEAL | – | 97.7% |
| SWE-Bench Pro V2 HARD | Scale AI SEAL | – | 86.3% |
| ALE-Bench | Epoch AI Benchmarking Hub | 1150.3† | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | – | 13.1% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | – | 81.3% |
| Design Arena Website | Design Arena | 1221 | 1335 |
| FrontierCode 2.0 (AA) | Artificial Analysis | – | 25.4% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 57.3%† | 13.0%† |
| APEX-Agents | Mercor | – | 46.8% |
| Arena Agent | Arena (LMArena) | – | 2.5 |
| Terminal-Bench Hard | Artificial Analysis | 37.9% | – |
| τ²-Bench Telecom (AA) | Artificial Analysis | 93.0% | – |
| Terminal-Bench 2.1 (Vals) | Vals AI | 44.2% | – |
| MCP Atlas | Scale AI SEAL | – | 85.4% |
| HiL-Bench | Scale AI SEAL | 20.0%† | – |
| GDPval-AA v2.1 | Artificial Analysis | – | 1630 |
| τ³-Banking (AA) | Artificial Analysis | – | 40.4% |
| Analyst Agent (AA) | Artificial Analysis | – | 51.3% |
| APEX-Agents (AA) | Artificial Analysis | 14.2% | – |
| CyberBench | Vals AI | – | 72.7% |
| Finance Agent v2 | Vals AI | 28.5% | – |
| Finance Agent v2.1 | Vals AI | – | 67.5% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | – | 21.7% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 0.0% | – |
| Legal Research Bench | Vals AI | 13.9% | 46.6% |
| SREBench | Vals AI | – | 1.9% |
| Tax Agent Bench | Vals AI | 7.0% | 38.9% |
| Terminal-Bench 2.0 (Vals) | Vals AI | 40.5% | – |
| Terminal-Bench 4.0 (Vals) | Vals AI | – | 12.6% |
| Terminal-Bench Science | Vals AI | – | 2.9% |
| Arena Agent Task Completion | Arena (LMArena) | – | 3.3 |
| Vending-Bench 2 | Andon Labs | 4662.9† | – |
| AutomationBench | Artificial Analysis | – | 45.6% |
| GDP.pdf (AA) | Artificial Analysis | – | 25.2% |
| EnterpriseOps-Gym | Artificial Analysis | – | 46.8% |
| AA-Briefcase v1.1 | Artificial Analysis | – | 1527 |
| Design Arena Agentic | Design Arena | 1093 | 1315 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | – | 5.7% |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 44.9% | – |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 17.1% | – |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 92.2% | – |
| ProofBench | Vals AI | 14.0% | – |
| AIME (Vals) | Vals AI | 96.5% | – |
| ArXivMath | MathArena | – | 31.9% |
| BrokenArXiv | MathArena | – | 63.2% |
| Arena Maths | Arena (LMArena) | 1464 | – |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 30.2% | – |
| AA-Omniscience v1.1 | Artificial Analysis | – | 16.1 |
| MMLU-Pro | Vals AI | 86.3% | – |
| LegalBench | Vals AI | 77.7% | – |
| CorpFin | Vals AI | 63.7% | – |
| TaxEval | Vals AI | 74.1% | – |
| MedQA | Vals AI | 94.5% | – |
| BioMysteryBench | Vals AI | – | 70.0% |
| CaseLaw v2 | Vals AI | 54.5% | – |
| Excel Modeling Benchmark | Vals AI | 11.7% | – |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 69.5% |
| MedCode | Vals AI | 32.2% | – |
| MedScribe | Vals AI | 63.4% | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | – | 33.8% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | – | 73.3% |
| Global-MMLU-Lite (AA) | Artificial Analysis | 89.5% | – |
| Instruction following | |||
| IFBench | Artificial Analysis | 81.2% | – |
| Arena Instruction Following | Arena (LMArena) | 1448 | 1410 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1472 | 1412 |
| EQ-Bench Creative Writing v3 | EQ-Bench | 1574† | – |
| Arena Creative Writing | Arena (LMArena) | 1445 | 1345 |
| Arena Multi-turn | Arena (LMArena) | 1479 | 1415 |
| Arena Longer Queries | Arena (LMArena) | 1466 | 1410 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1263 | – |
| MMMU-Pro | Artificial Analysis | 74.6% | 81.6% |
| MMMU-Pro (Vals) | Vals AI | 83.5% | – |
| MortgageTax | Vals AI | 45.4% | – |
| SAGE | Vals AI | – | 88.6% |
| Humanity's Sixth Sense | Scale AI SEAL | – | 33.3% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 0.0%† | – |
| VIALS-AA | Handshake AI | – | 42.9% |
| Code2Video-AA | Artificial Analysis | – | 1020 |
| AA-3D | Artificial Analysis | – | 1027 |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 69.0% | 83.3% |
| Arena Document | Arena (LMArena) | 1440 | – |
| MLCR | Artificial Analysis | – | 20.6% |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 152.3† | – |
| AA Intelligence Index v5.0.1 | Artificial Analysis | – | 30.6 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 25.7 | – |
| AA Progress Index v1 | Artificial Analysis | 281.9 | – |
| Vals Index | Vals AI | – | 56.1 |
| Vals Finance Index | Vals AI | – | 61.5% |
| Vals Science Index | Vals AI | – | 32.7% |
| Vals Multimodal Index | Vals AI | 39.1% | – |
| Vals Legal Index | Vals AI | – | 73.9% |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | – | 16.2% |
| DistressBench | Scale AI SEAL | – | 82.3% |
| Metadata | |||
| AA Openness Index | Artificial Analysis | – | 38.9 |
Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.
What changed
- Grok 4.20 (thinking): 1 score revision.
- Inkling 2 Small: 1 score revision; Inkling 2 Small cached input price moved from $0.060/M to $0.030/M (50% down).
- Inkling 2 Small: 1 score revision.
- Inkling 2 Small: BenchLeader Index 58.4 → 57.9.
- Inkling 2 Small: 2 new results (FORTRESS, Design Arena Agentic).
- Inkling 2 Small: 2 new results (MysteryMechanism, Vals Science Index); 1 score revision.
- Inkling 2 Small: 3 new results (SciCode, APEX-Agents, MLCR).
- Inkling 2 Small: 8 new results (Arena Text, Arena Hard Prompts, Arena Coding…); 2 score revisions.
- Inkling 2 Small: BenchLeader Index 57.5 → 58.4.
- Inkling 2 Small: 1 new result (Arena WebDev).
Related comparisons
- Grok 4.20 vs Gemini 4 Carbon
- Grok 4.20 vs GPT-6 Astra
- Inkling 2 Small vs Gemini 4 Carbon
- Inkling 2 Small vs Claude Opus 5.5
- Inkling 2 Small vs GPT-6 Astra
- Inkling 2 Small vs Claude Fable 5.1
- Grok 4.20 vs GPT-6.1 Sol
- Grok 4.20 vs Gemini 4 Argon
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Grok 4.20 vs Inkling 2 Small: questions
- Is Grok 4.20 better than Inkling 2 Small?
- Grok 4.20 (thinking) and Inkling 2 Small are neck and neck on quality (58.3 vs 57.9, within the ±5.4 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
- Grok 4.20 vs Inkling 2 Small: which is cheaper, and will that change?
- Inkling 2 Small is cheaper today: $0.475 against $1.56 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Grok 4.20 or Inkling 2 Small better for coding?
- Inkling 2 Small scores higher in coding (60 vs 53 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1376 for Grok 4.20 against 1647 for Inkling 2 Small.
- Which is faster, Grok 4.20 or Inkling 2 Small?
- Grok 4.20 streams faster (111 vs 100 output tokens per second, Artificial Analysis 7-day median).
- Is Grok 4.20 or Inkling 2 Small better for agents and tool use?
- Inkling 2 Small scores higher in agentic tasks (61 vs 51 on the category index, where 50 is average).
More questions (4)
- Which is better for long documents?
- Inkling 2 Small scores higher on long-context benchmarks (66 vs 59). Grok 4.20 accepts more context: 2,000,000 against 1,048,576 tokens for Inkling 2 Small.
- Which has lower latency?
- Grok 4.20 answers first: 12.1 s against 49.0 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Do independent benchmarks agree on Grok 4.20 vs Inkling 2 Small?
- Across the 16 benchmarks both have, Grok 4.20 leads 6, Inkling 2 Small leads 10. Every overall index that rates both puts Grok 4.20 ahead.
- When was Inkling 2 Small released?
- Inkling 2 Small was released on 5 Nov 2026.