vs
2:7
overall capability wins (0 ties across 9 scored). GPT-5.6 Sol leads overall.
- Reasoning1:1
- Coding0:4
- Arena1:0
- Composite indices0:2
- Pricing3:0
- Speed2:0
- Specs0:2
Verdict
GPT-5.6 Sol leads coding (widest gap: +170.3 on WebDev Elo); Gemini 3.1 Pro ranks higher on LMArena (+3 Elo); GPT-5.6 Sol offers larger context (1.1M vs 1M).
- GPT-5.6 Sol leads coding (widest gap: +170.3 on WebDev Elo)
- Gemini 3.1 Pro ranks higher on LMArena (+3 Elo)
- GPT-5.6 Sol offers larger context (1.1M vs 1M)
- GPT-5.6 Sol allows ~2.0x more max output (128K vs 66K)
- Gemini 3.1 Pro is ~1.8x cheaper on a blended token basis than GPT-5.6 Sol
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Gemini 3.1 Pro
- $5
- GPT-5.6 Sol
- $9
Gemini 3.1 Pro is about 1.8x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
OpenAI
Family
Gemini
GPT-5.x
License
Proprietary
Proprietary
Open weights
No
No
Release
Feb 19, 2026
Jul 9, 2026
Knowledge cutoff
-
2026-02-16
API / provider
OpenAI
Modalities
text, image, audio, video → text
text, image → text
Specs
Context window
1M
1.1M
Max output
66K
128K
Parameters
—
—
Pricing
Input $/1M
$2
$4
Output $/1M
$12
$20
Blended $/1M (3∶1)
$4.50
$8
Speed
tok/s
128
96
TTFT (s)
0.4
0.45
Reasoning
MMLU-Pro
—
—
GPQA Diamond
94.3
94.1
Humanity's Last Exam
44.4
49.5
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
59.4
AA-LCR v1.1
—
84.0
CritPt
—
32.3
MMMU-Pro
—
83.4
IFBench
—
72.7
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
80.6
—
SWE-bench Pro
54.2
64.6
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
70.7
88.0
Aider Polyglot
—
—
Terminal-Bench 3
—
34.6
BigCodeBench
—
—
SciCode
—
57.1
CursorBench
—
67.2
SWE-Rebench
54.9
62.3
NL2Repo-Bench
—
—
DeepSWE
—
72.7
WebDev Arena
1,447
1,617
Terminal-Bench 4.0
—
37.3
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,487
1,484
| Metric | Gemini 3.1 Pro | GPT-5.6 Sol | Delta |
|---|---|---|---|
| Identity | |||
| Organization | OpenAI | — | |
| Family | Gemini | GPT-5.x | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Feb 19, 2026 | Jul 9, 2026 | — |
| Knowledge cutoff | - | 2026-02-16 | — |
| API / provider | OpenAI | — | |
| Modalities | text, image, audio, video → text | text, image → text | — |
| Specs | |||
| Context window | 1M | 1.1M | B +50K |
| Max output | 66K | 128K | B +62K |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | $2 | $4 | A +$2 |
| Output $/1M | $12 | $20 | A +$8 |
| Blended $/1M (3∶1) | $4.50 | $8 | A +$3.50 |
| Speed | |||
| tok/s | 128 | 96 | A +32 |
| TTFT (s) | 0.4 | 0.45 | A +0.05 |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 94.3 | 94.1 | A +0.2 pts |
| Humanity's Last Exam | 44.4 | 49.5 | B +5.1 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 59.4 | — |
| AA-LCR v1.1 | — | 84.0 | — |
| CritPt | — | 32.3 | — |
| MMMU-Pro | — | 83.4 | — |
| IFBench | — | 72.7 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | 80.6 | — | — |
| SWE-bench Pro | 54.2 | 64.6 | B +10.4 pts |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 70.7 | 88.0 | B +17.3 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | 34.6 | — |
| BigCodeBench | — | — | — |
| SciCode | — | 57.1 | — |
| CursorBench | — | 67.2 | — |
| SWE-Rebench | 54.9 | 62.3 | B +7.4 pts |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | 72.7 | — |
| WebDev Arena | 1,447 | 1,617 | B +170 Elo |
| Terminal-Bench 4.0 | — | 37.3 | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,487 | 1,484 | A +3 Elo |
Gemini 3.1 Pro
- Context
- 1M / 66K out
- Parameters
- —
- Price
- $2 / $12
- Speed
- 128 tok/s · 0.4s TTFT
- Modalities
- text, image, audio, video → text
- License
- Proprietary
Google's current Pro flagship with native multimodality and standout GPQA / ARC-AGI accuracy.
GPT-5.6 Sol
- Context
- 1.1M / 128K out
- Parameters
- —
- Price
- $4 / $20
- Speed
- 96 tok/s · 0.45s TTFT
- Modalities
- text, image → text
- License
- Proprietary
OpenAI's July 2026 flagship reasoning tier, tuned for hard STEM and agentic work; ChatGPT's top API SKU in the 5.6 family. Promotional Standard rates are $4/$20 (through at least Nov 21, 2026). Fast mode (2× price) delivers up to 2.5× Standard speed, and the Cerebras-powered Ultrafast preview reaches up to 750 tok/s.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.