MFX · E-MOUNT ISO 1600 SHTR 1/800 APT f2.8 DRIVE ISO C1 C2 C3 α0 SONY
Sony Alpha Series  ·  Medium Format Flagship  ·  2025
α0
Beyond medium format.
Beyond what was possible.
~201MP
Resolution
30fps
Lossless RAW
−20 EV
Electronic ND
6-layer
Hexa BSI CMOS
≥16 stops
Dynamic range
0.6nm
Process node
01  ·  Sensor
The world's first
hexa-spectral
medium format sensor.

Conventional image sensors interpolate color. They guess. The α0's hexa-stacked BSI CMOS architecture captures six discrete spectral channels — Red, Green, Blue, Magenta, Cyan, Yellow — simultaneously in a single exposure. Every photon is measured, not estimated. The result is a color accuracy and a dynamic range that no Bayer-pattern sensor, regardless of size or resolution, can approach.

Built by Sony Semiconductor on their next-generation stacked architecture, the sensor integrates 32GB of on-die Intel ZAM beneath the logic layer — enabling zero-compression frame buffering at 30fps full resolution.

Sensor key figures
Resolution201MP  ·  16,384 × 12,288 px
Sensor size~54 × 40.5mm (MF standard)
ArchitectureHexa-stacked BSI CMOS, 6 photosite layers
Pixel pitch3.3µm per photosite per layer
Dynamic range≥16.4 stops measured (ISO 64)
On-die buffer32GB stacked Intel ZAM, 512 GB/s
Color depth16-bit native per channel (96-bit raw)
Native ISO64 – 204800
Fig. 01  ·  Hexa-stacked BSI CMOS — photosite cross-section
PHOTONS µLens RED ~700nm R GREEN ~550nm G BLUE ~450nm B MAGENTA R+B M CYAN G+B C YELLOW R+G Y LOGIC / ROUTING LAYER L 32GB STACKED INTEL ZAM SILICON SUBSTRATE ~4.2µm STACK DEPTH 700nm 380nm FULL VISIBLE SPECTRUM 16,384 × 12,288 px ~201MP per layer 6-ch simultaneous capture
Color accuracy
ΔE <0.4
Six-channel color reconstruction achieves a color error below ΔE 0.4 under any illuminant — a figure previously only achievable with laboratory spectrophotometers. Skin tones, plant chlorophyll, and atmospheric haze are reproduced with scientific precision.
Metamerism index
MI 0.02
The hexa-layer architecture virtually eliminates metameric failure — the phenomenon where two colors appear identical under one light source but different under another. Measured at MI 0.02, compared to MI 1.8 for the best Bayer sensors.
Dynamic range
16.4 stops
Complementary channel pairs (M, C, Y) extend effective dynamic range by capturing detail in spectral regions where primary channels saturate. The result: 16.4 stops measured at ISO 64, with full tone linearity from deepest shadow to specular highlight.
Design  ·  Top Deck
Always-on.
Always present.

The α0's top deck is built around the principle that your hands should never leave the camera to check a setting. The always-on XCDLED status display shows ISO, shutter speed, and aperture in a persistent readout — legible in direct sunlight at up to 2,000 cd/m², invisible in pitch darkness. It does not sleep. It does not dim. It is simply always there. The panel is a single-stack Z-LED XCDLED with full RGB output and MEMSLEV per-pixel gating for true-infinite contrast — no backlight dimming zones needed at this scale.

Flanking the LCD: two dedicated metal dials — the Drive Mode selector and the ISO dial — both with firm, indexed detents. No touchscreen menus for these. No deep dives into a settings hierarchy. You dial in, you shoot.

Top XCDLEDAlways-on XCDLED  ·  320×80px  ·  2,000 nit peak
Display contentISO  ·  Shutter  ·  Aperture  ·  Battery  ·  Frames
Drive dial8-position indexed metal  ·  Single / Cont. H/L / Bracketing
ISO dial7-position indexed  ·  64 / 100 / 400 / 800 / 3200 / Auto / Ext.
Shutter buttonThreaded release socket  ·  two-stage
Front C buttonsC1 / C2 / C3  ·  fully assignable
Top deck — always-on XCDLED status display
EVF ISO 1600 SHTR 1/800 APT f2.8 ▮▮▮▮ 83% CF-A: 214GB ND −5 ALWAYS-ON XCDLED 320×80 · 2,000 nit · never dims ▲ CH CL S BKT WBB DRV DRIVE DIAL 8-pos indexed metal 64 100 400 3200 AUTO ISO ISO DIAL 7-pos indexed SHTR M Av Tv P MODE always on · never interrupted
02  ·  MFX Optics System
Every E-mount lens.
Medium format reach.

The α0 introduces the world's first built-in motorized optical expander — the MFX Element Group — a nine-element telecentric design that projects the 43mm image circle of any Sony FE lens onto the α0's 54mm medium format sensor plane. No vignetting. No corner softness. The full native sharpness and color correction of your existing glass, expanded to fill a sensor that sees everything.

Fused into the same optical path: a liquid crystal variable neutral density filter tunable from 0 to −20 stops in 1/3-stop increments. No physical glass movement. No color cast. No flicker at any frame rate. Electronically neutral across the entire visible spectrum.

Electronic ND — 0 to −20 stops
0 −5 −10 −15 −20
Clear
full pass
−5 EV
overcast ext.
−10 EV
bright sun f/1.4
−15 EV
solar / long exp.
−20 EV
~ND1,000,000
MFX element group9 elements, 7 groups — telecentric design
Input image circle≥43mm (full Sony FE coverage)
Output image circle54mm medium format
Teleconverter range1.0× – 2.0× motorized, 0.1× steps
Electronic ND0 to −20 stops, 1/3-stop increments
ND color shiftΔE <0.1 (electronically corrected)
ND flicker frequencyZero — continuous LC biasing
TC resolution lossNone — optical magnification only
Fig. 02  ·  MFX optical expander + VND element cross-section
43mm FF VND LC 0–−20EV L1 MOTO R L2 L3 L4 L5 L6 L7 L8-L9 54mm MF 1.0× – 2.0× TC 9-ELEMENT TELECENTRIC DESIGN SONY × ZEISS CO-OPTIMIZED COATINGS 0 EV −10 EV −20 EV ELECTRONIC ND RANGE
Design  ·  Ergonomics
Built for
medium format demands.
Front view  ·  annotated
MFX · E-MOUNT C1 C2 C3 α0 SONY MFX MOUNT C1 / C2 / C3 FRONT FRONT ELEMENT 8.5-STOP IBIS  ·  ACTIVE THROUGH MFX EXPANDER works with all E-mount lenses · full MF sensor coverage
Rear view  ·  annotated
9.44M OLED 90°↑ 1/800 f2.8 ISO1600 RAW+HEIF 30fps CF-A 4-AXIS TILT MENU PLAY DEL AF-JOY CMD ISO 1600 SHTR 1/800 CF TYPE-B ×2 TILTABLE EVF  ·  ±90° UP  ·  45° DOWN 9.44M dot OLED · 240fps refresh
03  ·  Processing Engine
SONY × INTEL
Co-designed silicon  ·  Exclusive to α0
Engineered at the boundary
of what silicon can do.

The α0's processing system is the product of a three-year joint silicon engineering program between Sony Imaging Products and Intel's Custom Foundry division. Sony contributed the Bionz-XR imaging pipeline architecture — the most advanced ISP Sony has ever designed. Intel contributed the QuickSync hardware encode engine, the AIBOOST NPU architecture, and the 06E process node itself. The result: the QuickSync Bionz-XR — a processing block that has no equivalent in any camera ever made.

Four QuickSync Bionz-XR cores run in parallel, each assigned to a sensor quadrant. Two dedicated AIBOOST NPU cores run the entire AI system — completely isolated from the imaging pipeline. The SoC is built on Intel's proprietary 0.6nm 06E efficient node — the second most advanced process node on earth, behind only Intel's own 0.4nm 04A performance node.

Process node context
Intel 04A  ·  performance node
0.4nm
04A — world's most advanced
Reserved for Intel's own x86 CPUs and discrete GPUs. Highest transistor density ever achieved in volume silicon manufacturing.
Intel 06E  ·  efficiency node  ·  α0 SoC
0.6nm
06E — world's #2 node
Optimized for sustained high-throughput workloads at minimal thermal output. Ideal for cameras where the ISP must run continuously without throttling.
Context
For reference: TSMC N3 (3nm) and Samsung 3GAE (3nm) are the leading commercially available foundry nodes. Intel's 06E at 0.6nm effective gate pitch represents a generational leap beyond what any third-party foundry currently ships.
Fig. 03  ·  QuickSync Bionz-XR core — internal architecture illustration
intel  06E sony  bxr BIONZ-XR  ·  SONY ISP 6-CH SPECTRAL DEMOSAIC RGBMCY → unified image space ADAPTIVE NR ENGINE per-channel 3D processing TONE PIPELINE S-Log4 / Hexa RAW curves 6-CH COLOR SCIENCE ΔE <0.4  ·  BT.2020 native LENS CORR distortion / CA HDR MERGE multi-frame align OUTPUT FORMATTER QUICKSYNC  ·  INTEL ENCODE RAW ENCODE ENGINE lossless / lossy RAW · 30fps HEIF / AVIF ENGINE 120fps HEIF 16K hardware VIDEO CODEC BLOCK 8K60 XAVC-HS  ·  ProRes RAW THROUGHPUT CTRL ~90 GB/s sustained encode STORE MGR CFexpress ctrl PCIe 4.0 ×4 lanes OUTPUT BUS INTERFACE SHARED DATA BUS ZAM INTERFACE  ·  128 GB/s  ·  ZERO-LATENCY ADDRESSABLE MEMORY L1/L2 ISP CACHE  ·  48MB L1/L2 ENCODE CACHE  ·  32MB INTEL 0.6nm 06E EFFICIENT NODE  ·  SONY × INTEL CO-DESIGN Single core  ·  42 TOPS imaging  ·  Quad cluster  ·  168 TOPS total Plus 2× AIBOOST NPU  ·  96 TOPS AI inference  ·  264 TOPS SoC aggregate
Processing performance
264 TOPS aggregate. Nothing close exists.
QS Bionz-XR (×4)
168 TOPS
AIBOOST NPU (×2)
96 TOPS
SoC Total
264 TOPS
Prev. gen A1 II
~38 TOPS
Apple M4 (neural)
38 TOPS
04  ·  AI Imaging System
The most advanced
subject detection system
in any camera.

Two dedicated AIBOOST NPU cores — completely isolated from the imaging pipeline — run the α0's AI system with 96 TOPS of combined inference performance. This is not an evolution of existing AF AI. The computational headroom provided by Intel's 06E node enables a fundamentally new approach: rather than classifying subjects by category, the α0 builds a continuous volumetric motion model of every detected entity in real time.

The result is an autofocus system that does not merely track where a subject is — it predicts where the subject will be, accounting for limb occlusion, body rotation, parallax shift from the photographer's own movement, and even the aerodynamic drag model of a subject moving at speed. At 30fps RAW, the lens begins driving to the predicted next position before the current frame even reaches the sensor.

AIBOOST NPU — inference tasks
Architecture2× AIBOOST NPU  ·  48 TOPS each
IsolationFully isolated pipeline — zero ISP cycle sharing
AI model precisionINT4 / INT8 mixed precision inference
Subject taxonomy247 distinct subject classes
Prediction horizon±180ms motion vector projection
Skeleton resolution132-point full-body pose estimate
Eye detectionBilateral  ·  corneal micro-reflex lock
Vehicle detectionWheel plane tracking, chassis pose estimation
Bird/wildlifeWing-phase AF  ·  feather edge sharpness target
Latency0.4ms detection → lens drive signal
Fig. 04  ·  AIBOOST AI pipeline — subject detection architecture
AIBOOST NPU COMPLEX  ·  2 CORES  ·  96 TOPS LIVE SENSOR FEED  ·  240fps AF CROP EXTRACT independent high-fps window from full sensor SPATIAL SCENE TOKENIZER divides frame into 16×16 semantic patches  ·  INT4 MULTI-CLASS SUBJECT CLASSIFIER 247 subject classes  ·  6-channel spectral input human · animal · vehicle · aircraft · sports · macro 132-PT POSE SKELETON full body · hands · face 132 landmarks tracked STEREO DEPTH ESTIMATE PDAF + AI depth fusion sub-mm depth resolution CORNEAL MICRO-REFLEX LOCK bilateral eye tracking  ·  0.2px corneal centroid maintains lock through glasses  ·  dark irises  ·  occlusion PREDICTIVE MOTION VECTOR ENGINE Bayesian trajectory model  ·  ±180ms lookahead compensates photographer body motion + subject physics aerodynamic drag  ·  limb occlusion prediction  ·  parallax model WING-PHASE AF avian wing-beat synchronize peak-extension predict VEHICLE TRACK wheel plane estimation chassis yaw / speed model LENS DRIVE SIGNAL  ·  0.4ms LATENCY pre-positions to predicted next frame before shutter TOTAL AF PIPELINE  ·  DETECT TO DRIVE  ·  0.4ms fully async from imaging pipeline INT4 inference  ·  96 TOPS  ·  48 TOPS per core
Subject classes
247
247 distinct subject taxonomies trained on 2.4 billion annotated frames. Humans, animals, vehicles, aircraft, macro subjects, sporting equipment, musical instruments. Anything you point at, the α0 understands.
Skeleton tracking
132pt
132-point full-body pose skeleton maintained at 240fps through limb occlusion, extreme body angles, and multi-subject overlap. The system knows where every joint is — even when the lens cannot see it.
Prediction lookahead
±180ms
Bayesian trajectory prediction projects subject position ±180ms into the future. The α0 begins driving the lens to the next position before the current frame finishes reading out of the sensor.
Lens drive latency
0.4ms
From detection pixel to lens actuator signal: 0.4 milliseconds. Running on isolated AIBOOST silicon with direct hardware interrupt to the lens drive controller — no OS scheduling, no pipeline stalls.
05  ·  Autofocus System
Phase detection
reimagined for
201 megapixels.

Phase detection autofocus has been Sony's foundation since the α7 series. The α0 does not abandon that foundation — it multiplies it by six. Every photosite on all six spectral layers of the hexa-stacked sensor participates in phase detection simultaneously. Where previous Sony cameras used a subset of dedicated PDAF pixels interspersed within a Bayer array, the α0's architecture means every single pixel on every single channel contributes phase disparity data in every exposure. There are no "AF pixels" and "image pixels." They are the same pixel.

This is Hexa-PDAF — a phase detection architecture that has never existed before, because it requires a sensor that has never existed before. Six independent phase disparity maps, one per spectral channel, are fused by the AIBOOST NPU into a single volumetric depth field across the full 201MP frame. The result: autofocus that is simultaneously the most coverage-dense, the most depth-accurate, and the most contrast-independent AF system ever placed in a camera.

Focus acquisition: 0.02 seconds. Focus tracking: continuous at 30fps RAW, zero frame loss. Low-light PDAF: down to −7 EV — four stops darker than any previous Sony Alpha. In complete darkness, the Spectral Phase Contrast engine uses the differential phase response between the R and Cy channels — wavelengths separated by 250nm — to extract depth information that is invisible to any conventional monochrome or bayer PDAF system.

Hexa-PDAF — coverage map
COVERAGE 100% FRAME · 6 LAYERS
Each dot = one PDAF-capable photosite cluster ~201M PDAF points across 6 channels
AF system specifications
PDAF architectureHexa-PDAF — all photosites, all 6 channels
PDAF coverage100% frame × 100% height
PDAF points~1.26 billion phase pairs (201M × 6ch × dual)
Focus acquisition0.02s from anywhere in frame
PDAF low-light limit−7 EV (Spectral Phase Contrast mode)
Spectral Phase ContrastR/Cy channel differential · 250nm wavelength gap
Touch AFReal-time · any point in live view
Focus modesSingle-shot AF · Continuous AF · Predictive · Manual
Zone modesWide · Zone · Flexible Spot · Tracking · Pinpoint
AF+MFHalf-press AF lock · Focus hold button · Focus map overlay
Focus peaking6-channel spectral peaking · 5 colour options
MF AssistPhase Map live display · depth contour lines
Fig. 05  ·  Hexa-PDAF depth field reconstruction
SCENE / SUBJECT R G B M C Y 6 INDEPENDENT PHASE DISPARITY MAPS AIBOOST NPU PHASE FUSION 6-channel disparity merge · Bayesian depth weighting · outlier rejection VOLUMETRIC DEPTH FIELD MAP ∞ FAR SUBJECT PLANE NEAR SPECTRAL PHASE CONTRAST  ·  −7 EV R channel ~700nm vs Cy channel ~480nm · 220nm spectral separation R 700nm Cy 480nm phase Δ = depth LENS DRIVE OUTPUT 0.02s acquisition · 0.4ms continuous drive latency NO DEDICATED AF PIXELS. EVERY PIXEL IS AN AF PIXEL. ~1.26 billion phase pairs active per exposure
PDAF points
~1.26B
~1.26 billion phase-detection pairs across the full frame — roughly 30× the PDAF density of the α1 II. Every pixel on every channel contributes. No coverage gaps, no corner falloff.
Acquisition speed
0.02s
Focus acquisition from anywhere in the frame: 0.02 seconds. Enabled by the density of Hexa-PDAF data and the AIBOOST NPU processing all six disparity maps in parallel with dedicated silicon.
Low-light limit
−7 EV
Phase detection remains reliable down to −7 EV — starlight conditions. The Spectral Phase Contrast engine extracts depth from the wavelength-separated R and Cy channel phase offset even when scene contrast is effectively zero.
Frame coverage
100%
Full 100% frame width and height PDAF coverage, including extreme corners. No corner zones where contrast-detect fallback is required. Corner subjects lock identically to center subjects.
Autofocus  ·  Focus modes
Every mode.
Mastered.

Focus mode design on the α0 follows the same philosophy as the rest of the camera: the technology should disappear. AF-C tracking through a 30fps RAW burst at 201MP never drops a frame. Wide Area AF doesn't just find the nearest subject — it identifies all subjects simultaneously, assigns a confidence weighting from the AIBOOST subject classifier, and prioritizes based on your historical shooting pattern. The camera learns which eye you prefer to focus on when photographing a couple. It remembers.

Focus mode reference
AF-SSingle-shot · locks on half-press · confirms with audible + visual
AF-CContinuous · predictive · 30fps RAW zero frame drop guarantee
AF-AAutomatic — switches S↔C based on subject motion detection
DMFDirect MF — AF acquire then manual fine-tune without re-engaging
Wide AreaAll 201MP, all subjects, confidence-ranked selection
ZoneUser-defined zone, any shape, up to 32MP zone area
Flexible SpotS / M / L spot size · free placement anywhere in frame
Expand Flex SpotFlexible spot + adjacent zone fallback if subject exits spot
TrackingTouch-to-track · AIBOOST-accelerated · 132-pt pose lock
Pinpoint AFSub-pixel accuracy · peaking + magnification combined
Pre-AFContinuous passive focus estimation even without half-press
Focus holdAssignable to any front C button or back AF-ON
Focus zone coverage — Wide Area AF
HUMAN · PRIMARY eye lock · L eye HUMAN · 2 ZONE user-defined AF-C  ·  WIDE AREA  ·  HUMAN 30fps RAW 100% FRAME HEXA-PDAF COVERAGE Primary subject (eye-tracked) Secondary subject (tracked) User-defined zone Eye detection lock point FOCUS ACQUISITION TIMELINE 0.02s acquire 0.4ms per-frame drive pulse  ·  30fps
06  ·  XCDLED Display System
The first camera
with XCDLED
everywhere.

Sony's XCDLED is not an incremental improvement over OLED or miniLED. It is a fundamentally new display architecture — a tandem dual-layer, dual-mode GaN–QD–MEMS light-valve panel that surpasses CRT motion clarity, OLED contrast, and miniLED peak brightness simultaneously. The α0 is the first camera to deploy XCDLED across all three of its displays: the rear LCD, the electronic viewfinder, and the top-deck status display. Three different panel sizes. Three different refresh targets. The rear LCD and EVF run the full tandem dual-stack architecture. The top status display uses a smaller, simpler single-stack Z-LED XCDLED — no TANDEM, no dimming zones, just Z-LED with RGB output and MEMSLEV gating for perfect blacks. All three representing the state of the art in their respective class.

Rear LCD  ·  3.0-type
4K XCDLED.
120 Hz. Per pixel.

The rear monitor is a 3.0-type XCDLED panel running at native 4K (3,840 × 2,160) — 8,294,400 dots. Where a conventional rear LCD averages roughly 2 million dots at best, the α0's rear panel resolves every pixel of a 4K source without interpolation. The XCDLED architecture layers a UV-pumped MCY quantum-dot stack beneath a blue-pumped RGB stack, gated by a 4K-resolution MEMS light-valve layer. Each pixel operates at sub-0.4 ms switching, enabling the 120 Hz refresh that eliminates sample-and-hold blur during live view — even at 30fps RAW throughput.

Peak brightness reaches 21,000 cd/m² at the highlight level. SDR baseline — the mode you use for menu navigation in sunlight — runs at a stable 450 cd/m² regardless of ambient light. In direct afternoon sun, the rear panel reads as clearly as paper.

Rear LCD — XCDLED panel specifications
Panel typeSony XCDLED — tandem GaN–QD–MEMSLV
Size3.0-type (7.62 cm diagonal)
Resolution3,840 × 2,160 dots  ·  8,294,400 total
Pixel density~732 ppi
Refresh rate120 Hz  ·  MEMSLV switching <0.4 ms
PrimariesDual-stack RGB + MCY  ·  6 spectral primaries
SDR brightness250–450 cd/m² (user-selectable)
HDR peak (highlight)21,000 cd/m²
Dual-stack peak~38,000 cd/m² (theoretical flashbang)
Contrast ratioTrue infinite — per-pixel MEMSLV black
Color accuracyΔE <0.05  ·  exceeds Rec.2020
GamutExceeds DCI-P3, Rec.2020 & human gamut boundaries
Dimming zones128 × 64 per GaN stack (16,384 zones)
Touch10-point capacitive  ·  glove & rain mode
Tilt4-axis — up 90° / down 45° / left 90° / right 20°
Fig. 06  ·  XCDLED rear LCD — panel layer cross-section
GaN BACKLIGHT (UV + BLUE) UV ~405nm MCY QD layer MCY COLLIMATION LENS COL BLUE ~450nm RGB QD+pass MCY↑ MEMS LIGHT VALVE (MEMSLV) 3,840 × 2,160 gates  ·  <0.4ms switching MEMS LV gate 120Hz MONOLITHIC NARROW LENS L7 OPTICS OPTIC IMAGE OUTPUT → VIEWER Peak: 21,000 cd/m² · SDR: 250–450 cd/m² · Dual-stack: ~38,000 cd/m² 3,840 × 2,160 dots · 8,294,400 total · ~732 ppi · 120 Hz · ΔE <0.05
6-Primary Efficiency
MCY + RGB
Operating as a 6-primary RGBMCY system, each primary is rendered by its closest native QD layer — M, C, Y from the UV-pumped stack; R, G, B from the blue-pumped stack. The unused layer powers down completely. Maximum efficiency, minimum heat, zero additive error.
Dual-stack peak mode
38Kcd/m²
Both GaN stacks fire simultaneously. Each contributes ~12K cd/m² at the panel, combining through the MEMSLV gate for a theoretical ~38,000 cd/m² burst peak. Used for HDR highlight bursts in extreme direct sunlight or sun-facing outdoor shooting.
Electronic Viewfinder  ·  0.64-type
8K XCDLED EVF.
240 Hz. Zero persistence.

The electronic viewfinder deploys the same tandem GaN–QD–MEMSLV XCDLED architecture as the rear LCD — in a 0.64-type microdisplay format — at 8K resolution (7,680 × 4,320) and 240 Hz refresh. At 240 Hz with sub-0.4 ms MEMSLV gating, the EVF's moving-image persistence is lower than a CRT phosphor at the same brightness. There is no sample-and-hold blur. There is no motion judder. A subject moving at 80 km/h against a complex background resolves as a sharp edge in the EVF — not a smear — at any shutter speed.

The 8K panel renders 33,177,600 dots at ~1,400 ppi — a dot density that exceeds the resolving power of the human eye at normal eyepiece distances. What the photographer sees in the α0 EVF is not a screen approximation of the scene. It is the scene. The on-panel Meridian controller IC processes the AV2-compressed display stream independently from the imaging ISP, ensuring the viewfinder never stutters regardless of RAW throughput.

EVF peak brightness reaches 10,000 cd/m² — sufficient for clear viewing in direct sunlight. The tiltable housing allows ±90° vertical and ±45° horizontal articulation, enabling shooting from the hip, overhead, or at ground level without removing the eye from the finder.

EVF — XCDLED microdisplay specifications
Panel typeSony XCDLED — tandem GaN–QD–MEMSLV
Size0.64-type (1.63 cm diagonal, microdisplay)
Resolution7,680 × 4,320 dots  ·  33,177,600 total
Pixel density~1,400 ppi
Refresh rate240 Hz  ·  MEMSLV switching <0.4 ms
PersistenceSub-CRT phosphor decay at all brightness levels
PrimariesDual-stack RGB + MCY  ·  6 spectral primaries
Peak brightness10,000 cd/m²
Contrast ratioTrue infinite — per-pixel MEMSLV black
Color accuracyΔE <0.05  ·  exceeds Rec.2020
Magnification0.90× (35mm equiv.)
Eye relief23mm
Tilt±90° vertical  ·  ±45° horizontal
EVF renderingMeridian Display Controller — async from ISP
Diopter−4.0 to +3.0 m⁻¹  ·  motorised
Fig. 07  ·  XCDLED EVF — microdisplay cross-section + viewfinder render
XCDLED MICRODISPLAY · 0.64-type 1/800 f2.8 ISO1600 7,680 × 4,320 dots · ~1,400 ppi EYEPIECE OPTICS 0.90× 90°↑ 90°↓ XCDLED LAYER ARCHITECTURE · EVF UV GaN HOLLOW CYLINDER ARRAY · ~405nm L1 MCY QD CONVERSION · MAGENTA / CYAN / YELLOW L2 MCY COLLIMATION LENS ARRAY L3 BLUE GaN Z-LED ARRAY · ~450nm L4 RGB QD CONVERSION · RED / GREEN / BLUE + MCY PASS L5 MEMSLV — 8K MEMS LIGHT VALVE <0.4ms · 240Hz · per-pixel · true-infinite contrast L6 SINGLE OUTPUT OPTICS L7 LENS ARRAY L7 TO EYEPIECE OPTICS → EYE PERSISTENCE COMPARISON CRT OLED XCDLED α0 EVF ~1ms ~2ms <0.4ms ✓ persistence → (shorter bar = less blur)
EVF resolution
33.2M dot
7,680 × 4,320 native 8K. At ~1,400 ppi the dot structure disappears entirely at normal eye-relief distances. You see the scene, not the display.
Refresh
240Hz
240 Hz with sub-0.4 ms MEMSLV gating. Persistence drops below CRT phosphor decay — the only display technology to achieve this at camera-class brightness.
Rear LCD
8.3M dot
3,840 × 2,160 native 4K on the rear panel — more dots than most broadcast monitors. 120 Hz XCDLED. True-infinite contrast. 21,000 cd/m² peak.
Peak brightness
21Kcd/m²
Both panels run the same XCDLED core architecture. Both achieve 21,000 cd/m² HDR peak. Both deliver true-infinite per-pixel contrast via the MEMSLV gate layer.
07  ·  Full Specifications
Complete technical
specification.
Sensor system
FormatMedium format, ~54 × 40.5mm
ArchitectureHexa-stacked BSI CMOS
Resolution201,326,592 px  ·  16,384 × 12,288
Pixel pitch3.3µm per photosite per channel
Spectral layersR, G, B, Magenta, Cyan, Yellow
Color depth16-bit per channel (96-bit total RAW)
Dynamic range16.4 stops (ISO 64, measured)
Color accuracyΔE <0.4 all illuminants
Metamerism indexMI 0.02 (vs. MI 1.8+ Bayer)
On-die DRAM32GB stacked Intel ZAM · 512 GB/s
Native ISO range64 – 204800
Extended ISO32 / 409600
Color spacesACES / BT.2020 / DCI-P3 / AdobeRGB
Optics & VND
Lens mountSony E-Mount · full FE compatibility
MFX expander9 elements · 7 groups · telecentric
FF image circle in≥43mm coverage
MF image circle out54mm at sensor plane
Teleconverter1.0× – 2.0× motorized · 0.1× steps
Electronic ND0 to −20 EV · 1/3-stop steps · stepless
ND color neutralityΔE <0.1 (electronically corrected)
ND flickerNone · continuous LC bias voltage
ND spectral range380–750nm uniform attenuation
Autofocus
PDAF architectureHexa-PDAF · all photosites · 6 channels
PDAF coverage100% frame width × 100% height
PDAF pairs~1.26 billion phase pairs per exposure
Focus acquisition0.02s from anywhere in frame
Low-light PDAF−7 EV · Spectral Phase Contrast engine
AF-C at 30fps RAWZero frame drop · continuous drive guarantee
Subject classes247 AI subject taxonomies
Pose tracking132-point skeleton · 240fps
Eye detectionCorneal micro-reflex · bilateral
AF predictionBayesian ±180ms trajectory model
AF latency0.4ms detect → lens drive
Focus peaking6-channel spectral peaking · 5 colour options
Processor & AI
SoCQuickSync Bionz-XR · Sony × Intel
Process nodeIntel 0.6nm 06E efficient architecture
ISP cores4× QuickSync Bionz-XR · 42 TOPS each
ISP total TOPS168 TOPS imaging throughput
AI NPU cores2× AIBOOST · 48 TOPS each
AI total TOPS96 TOPS inference · fully isolated
SoC aggregate264 TOPS combined
ZAM bufferZero-latency addressable · 128 GB/s
Display system
EVF typeSony XCDLED microdisplay · GaN–QD–MEMSLV
EVF resolution7,680 × 4,320 dots  ·  33,177,600 total
EVF dot density~1,400 ppi
EVF refresh240 Hz · <0.4ms MEMSLV switching
EVF peak brightness10,000 cd/m²
EVF contrastTrue infinite · per-pixel MEMSLV gate
EVF magnification0.90× · 23mm eye relief
EVF tilt±90° vertical · ±45° horizontal
Rear LCD typeSony XCDLED · GaN–QD–MEMSLV · 3.0-type
Rear LCD resolution3,840 × 2,160 dots  ·  8,294,400 total
Rear LCD density~732 ppi
Rear LCD refresh120 Hz · <0.4ms MEMSLV switching
Rear LCD peak21,000 cd/m² HDR · ~38,000 cd/m² dual-stack
Rear LCD SDR250–450 cd/m² (user-selectable)
Rear LCD contrastTrue infinite · per-pixel MEMSLV gate
Rear LCD colorΔE <0.05 · exceeds Rec.2020
Rear LCD touch10-point capacitive · glove + rain mode
Rear LCD tilt4-axis — 90° up / 45° down / 90° left / 20° right
Top displayAlways-on XCDLED  ·  320×80  ·  2,000 nit  ·  never dims
Capture & output
Full-res RAW30fps lossless · 60fps compressed
HEIF capture120fps full 16K resolution
Burst depthUnlimited (ZAM to CFexpress direct)
Video8K60 XAVC-HS · 4K240 · ProRes RAW
Log profilesS-Log4 · S-Cinetone · Hexa RAW
CFexpressDual Type B · PCIe 4.0 ×4 · ~14 GB/s
ConnectivityUSB4 · Wi-Fi 7 · Bluetooth 5.4
Ergonomicsα1 II form factor · deeper MF grip
BatteryNP-FZ100 compatible · 2× grip support
α0
Six layers of light. A sensor that measures, not guesses. A processor born from the most advanced foundry partnership in consumer electronics history. An AI that knows your subject's next move before you do. Phase detection on every pixel, across six simultaneous spectral channels, down to starlight. And two XCDLED panels — one for your eye, one for your hand — that redefine what a camera screen can be.

This is the α0. The only Alpha that can beat "the one."
Sony α0  ·  Beyond medium format  ·  Concept 2025
Not a real product  ·  All specifications are fictional