vs
1:1
overall capability wins (0 ties across 2 scored). Grok 4.5 leads overall.
- Pricing3:0
- Speed0:2
- Specs1:1
Verdict
Grok 4.20 offers ~2.0x larger context (1M vs 500K); Grok 4.5 allows ~2.0x more max output (128K vs 64K); Grok 4.20 is ~1.9x cheaper on a blended token basis than Grok 4.5.
- Grok 4.20 offers ~2.0x larger context (1M vs 500K)
- Grok 4.5 allows ~2.0x more max output (128K vs 64K)
- Grok 4.20 is ~1.9x cheaper on a blended token basis than Grok 4.5
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Grok 4.20
- $1.88
- Grok 4.5
- $3.50
Grok 4.20 is about 1.9x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
xAI
xAI
Family
Grok
Grok
License
Proprietary
Proprietary
Open weights
No
No
Release
May 20, 2026
Jul 9, 2026
Knowledge cutoff
-
-
API / provider
xAI
xAI
Modalities
text, image → text
text, image → text
Specs
Context window
1M
500K
Max output
64K
128K
Parameters
—
—
Pricing
Input $/1M
$1.25
$2
Output $/1M
$2.50
$6
Blended $/1M (3∶1)
$1.56
$3
Speed
tok/s
83
100
TTFT (s)
0.4
0.35
Reasoning
MMLU-Pro
—
—
GPQA Diamond
—
93.1
Humanity's Last Exam
—
42.7
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
51.6
AA-LCR v1.1
—
79.3
CritPt
—
15.4
MMMU-Pro
—
80.4
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
—
81.6
Aider Polyglot
—
—
Terminal-Bench 3
—
15.7
BigCodeBench
—
—
SciCode
—
55.0
CursorBench
—
—
SWE-Rebench
—
63.8
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
1,555
Terminal-Bench 4.0
—
12.4
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
1,468
| Metric | Grok 4.20 | Grok 4.5 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | xAI | xAI | — |
| Family | Grok | Grok | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | May 20, 2026 | Jul 9, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | xAI | xAI | — |
| Modalities | text, image → text | text, image → text | — |
| Specs | |||
| Context window | 1M | 500K | A +500K |
| Max output | 64K | 128K | B +64K |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $1.25 | $2 | A +$0.75 |
| Output $/1M | $2.50 | $6 | A +$3.50 |
| Blended $/1M (3∶1) | $1.56 | $3 | A +$1.44 |
| Speed | |||
| tok/s | 83 | 100 | B +17 |
| TTFT (s) | 0.4 | 0.35 | B +0.05 |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | — | 93.1 | — |
| Humanity's Last Exam | — | 42.7 | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 51.6 | — |
| AA-LCR v1.1 | — | 79.3 | — |
| CritPt | — | 15.4 | — |
| MMMU-Pro | — | 80.4 | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | 81.6 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | 15.7 | — |
| BigCodeBench | — | — | — |
| SciCode | — | 55.0 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | 63.8 | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | 1,555 | — |
| Terminal-Bench 4.0 | — | 12.4 | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | 1,468 | — |
Grok 4.20
- Context
- 1M / 64K out
- Parameters
- —
- Price
- $1.25 / $2.50
- Speed
- 83 tok/s · 0.4s TTFT
- Modalities
- text, image → text
- License
- Proprietary
High-speed Grok 4.x mid-tier with 1M context at $1.25/$2.50 (doubles above 200k prompt tokens); current API snapshot is grok-4.20-0309-reasoning.
Grok 4.5
- Context
- 500K / 128K out
- Parameters
- —
- Price
- $2 / $6
- Speed
- 100 tok/s · 0.35s TTFT
- Modalities
- text, image → text
- License
- Proprietary
xAI/SpaceXAI public API flagship (July 2026) - Cursor-trained coding focus at $2/$6. For the Cursor IDE jointly-trained SKU and Fast tier, see Cursor Grok 4.5.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.
These two models have no overlapping published benchmarks in our dataset. Compare specs and pricing instead.